diff --git "a/results.json" "b/results.json"
--- "a/results.json"
+++ "b/results.json"
@@ -620,7 +620,7 @@
       "family": "Indo-European",
       "flores_path": "nld_Latn",
       "fleurs_tag": "nl_nl",
-      "commonvoice_hours": 117.0,
+      "commonvoice_hours": 118.0,
       "commonvoice_locale": "nl",
       "in_benchmark": true
     },
@@ -1292,7 +1292,7 @@
       "family": "Indo-European",
       "flores_path": "cat_Latn",
       "fleurs_tag": "ca_es",
-      "commonvoice_hours": 2852.0,
+      "commonvoice_hours": 2851.0,
       "commonvoice_locale": "ca",
       "in_benchmark": true
     },
@@ -3560,7 +3560,7 @@
       "family": "Abkhaz-Adyge",
       "flores_path": null,
       "fleurs_tag": null,
-      "commonvoice_hours": 37.0,
+      "commonvoice_hours": 38.0,
       "commonvoice_locale": "kbd",
       "in_benchmark": false
     },
@@ -4616,7 +4616,7 @@
       "family": "Turkic",
       "flores_path": null,
       "fleurs_tag": null,
-      "commonvoice_hours": 13.0,
+      "commonvoice_hours": 14.0,
       "commonvoice_locale": "sah",
       "in_benchmark": false
     },
@@ -4976,7 +4976,7 @@
       "family": "Indo-European",
       "flores_path": null,
       "fleurs_tag": null,
-      "commonvoice_hours": 0.5,
+      "commonvoice_hours": 0.6,
       "commonvoice_locale": "kvx",
       "in_benchmark": false
     },
@@ -5432,7 +5432,7 @@
       "family": "Indo-European",
       "flores_path": null,
       "fleurs_tag": null,
-      "commonvoice_hours": 0.1,
+      "commonvoice_hours": 1.1,
       "commonvoice_locale": "gjk",
       "in_benchmark": false
     },
@@ -8429,3459 +8429,13011 @@
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.15815751066481462,
-      "sentence_nr": 1
+      "score": 0.4226799078177409,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5152611872266766,
-      "sentence_nr": 1
+      "score": 0.5651672709988255,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.07407154448063642,
-      "sentence_nr": 1
+      "score": 0.32406433662077544,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.43145434527321425,
-      "sentence_nr": 1
+      "score": 0.5243586266504104,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.12903696060775005,
-      "sentence_nr": 1
+      "score": 0.34633672321253084,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.456225988032654,
-      "sentence_nr": 1
+      "score": 0.5378805625051344,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.12369892692249995,
-      "sentence_nr": 1
+      "score": 0.5077888484472814,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.44549610902403686,
-      "sentence_nr": 1
+      "score": 0.6493197366069867,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.06647168102389285,
-      "sentence_nr": 1
+      "score": 0.4318843329340524,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.34350832619898364,
-      "sentence_nr": 1
+      "score": 0.6011096108554106,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.12560672881768975,
-      "sentence_nr": 1
+      "score": 0.3582301850807646,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.4969560260291519,
-      "sentence_nr": 1
+      "score": 0.5380305837807603,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.12422788549118892,
-      "sentence_nr": 1
+      "score": 0.3732667150787326,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.40222210564426,
-      "sentence_nr": 1
+      "score": 0.5674650482249737,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.09735981717515908,
-      "sentence_nr": 1
+      "score": 0.4641883721676649,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.35288934658906385,
-      "sentence_nr": 1
+      "score": 0.6403267149729506,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.08273178236238297,
-      "sentence_nr": 1
+      "score": 0.300740577257699,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.36399666460809255,
-      "sentence_nr": 1
+      "score": 0.5272774705181614,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.38870674200492367,
-      "sentence_nr": 2
+      "score": 0.3576035471132581,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6484380084879691,
-      "sentence_nr": 2
+      "score": 0.5426399702952437,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.3961285597009415,
-      "sentence_nr": 2
+      "score": 0.4422044705926463,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6148751441350505,
-      "sentence_nr": 2
+      "score": 0.6089032707320831,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.4923751299732868,
-      "sentence_nr": 2
+      "score": 0.3099603853356145,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6853756490381199,
-      "sentence_nr": 2
+      "score": 0.5209233176748354,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.11133996756497437,
-      "sentence_nr": 2
+      "score": 0.33210944907163426,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.4410280353998367,
-      "sentence_nr": 2
+      "score": 0.5289420578289948,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.17374951565433233,
-      "sentence_nr": 2
+      "score": 0.4331131003868224,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.45325597884524305,
-      "sentence_nr": 2
+      "score": 0.5898969623074624,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.17743299460161885,
-      "sentence_nr": 2
+      "score": 0.35580399268816465,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.43071271897416463,
-      "sentence_nr": 2
+      "score": 0.5392592206305507,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.26459538953931094,
-      "sentence_nr": 2
+      "score": 0.4475435253337274,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5272178908335121,
-      "sentence_nr": 2
+      "score": 0.5956867226653717,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.26801022984888695,
-      "sentence_nr": 2
+      "score": 0.5274220384037692,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5654883864995515,
-      "sentence_nr": 2
+      "score": 0.6765588140322357,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.21665407194210906,
-      "sentence_nr": 2
+      "score": 0.39317381456022266,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.4344921442639243,
-      "sentence_nr": 2
+      "score": 0.6026058740561834,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.38249626297768063,
-      "sentence_nr": 3
+      "score": 0.4166560818400039,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.40976234193505356,
-      "sentence_nr": 3
+      "score": 0.6515522498665886,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.7281051247089317,
-      "sentence_nr": 3
+      "score": 0.42734667499155,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7882997401328445,
-      "sentence_nr": 3
+      "score": 0.6397906518456509,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.5806197937310393,
-      "sentence_nr": 3
+      "score": 0.48930936408255293,
+      "sentence_nr": 0
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7346706700987636,
-      "sentence_nr": 3
+      "score": 0.699085629239476,
+      "sentence_nr": 0
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "en",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.08635800047213174,
-      "sentence_nr": 3
+      "score": 0.15815751066481462,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "en",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.218109371254876,
-      "sentence_nr": 3
+      "score": 0.5152611872266766,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "en",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.25552199116069907,
-      "sentence_nr": 3
+      "score": 0.07407154448063642,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "en",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3799133205289109,
-      "sentence_nr": 3
+      "score": 0.43145434527321425,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "en",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.23386786214190372,
-      "sentence_nr": 3
+      "score": 0.12903696060775005,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "en",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3682311523733465,
-      "sentence_nr": 3
+      "score": 0.456225988032654,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "zh",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.5642761727828352,
-      "sentence_nr": 3
-    },
+      "score": 0.12369892692249995,
+      "sentence_nr": 1
+    },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "zh",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6181373706707737,
-      "sentence_nr": 3
+      "score": 0.44549610902403686,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "zh",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.4093301993048525,
-      "sentence_nr": 3
+      "score": 0.06647168102389285,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "zh",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.512762518189388,
-      "sentence_nr": 3
+      "score": 0.34350832619898364,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "zh",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.6244631487487835,
-      "sentence_nr": 3
+      "score": 0.12560672881768975,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "zh",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6931369519059803,
-      "sentence_nr": 3
+      "score": 0.4969560260291519,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "hi",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.8003203203844999,
-      "sentence_nr": 4
+      "score": 0.12422788549118892,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "hi",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.9453478043428296,
-      "sentence_nr": 4
+      "score": 0.40222210564426,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "hi",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
-      "sentence_nr": 4
+      "score": 0.09735981717515908,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "hi",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
-      "sentence_nr": 4
+      "score": 0.35288934658906385,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "hi",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
-      "sentence_nr": 4
+      "score": 0.08273178236238297,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "hi",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
-      "sentence_nr": 4
+      "score": 0.36399666460809255,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.0,
-      "sentence_nr": 4
+      "score": 0.13714845589364738,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.629039349740581,
-      "sentence_nr": 4
+      "score": 0.45499281593451946,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.0,
-      "sentence_nr": 4
+      "score": 0.11564012893219777,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.629039349740581,
-      "sentence_nr": 4
+      "score": 0.44599783682350064,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.0,
-      "sentence_nr": 4
+      "score": 0.12601482779921785,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.629039349740581,
-      "sentence_nr": 4
+      "score": 0.43595665254608706,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.0,
-      "sentence_nr": 4
+      "score": 0.3026566818840519,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.49546288984677567,
-      "sentence_nr": 4
+      "score": 0.5945859352092411,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "bleu",
       "score": 0.0,
-      "sentence_nr": 4
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.43795381992037963,
-      "sentence_nr": 4
+      "score": 0.2521233582161207,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.0,
-      "sentence_nr": 4
+      "score": 0.40959087443621306,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5881561248602009,
-      "sentence_nr": 4
+      "score": 0.6348509381122925,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
-      "sentence_nr": 5
+      "score": 0.07793031063789554,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
-      "sentence_nr": 5
+      "score": 0.3700181221537743,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.2887138086538547,
-      "sentence_nr": 5
+      "score": 0.0867932999243575,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6342291345998248,
-      "sentence_nr": 5
+      "score": 0.4201964133235075,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
-      "sentence_nr": 5
+      "score": 0.08214106568089705,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
-      "sentence_nr": 5
+      "score": 0.3969463877642616,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.47410002229034043,
-      "sentence_nr": 5
+      "score": 0.0,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7538467008030766,
-      "sentence_nr": 5
+      "score": 0.41649654108052436,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.5087473540251254,
-      "sentence_nr": 5
+      "score": 0.0,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7647955332172516,
-      "sentence_nr": 5
+      "score": 0.3630576975795868,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.5087473540251254,
-      "sentence_nr": 5
+      "score": 0.0744904632040495,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7647955332172516,
-      "sentence_nr": 5
+      "score": 0.4111163205685468,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.0,
-      "sentence_nr": 5
+      "score": 0.08767210132815903,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.2828367156737383,
-      "sentence_nr": 5
+      "score": 0.40476518002703893,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.48181149445310956,
-      "sentence_nr": 5
+      "score": 0.08616711094288851,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7675828789334244,
-      "sentence_nr": 5
+      "score": 0.3696512763473903,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.5091224918749461,
-      "sentence_nr": 5
+      "score": 0.12894104034845807,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7829685247145245,
-      "sentence_nr": 5
+      "score": 0.4486368934849452,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
-      "sentence_nr": 6
+      "score": 0.14738500064905094,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
-      "sentence_nr": 6
+      "score": 0.4659728395318289,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.7511573912724299,
-      "sentence_nr": 6
+      "score": 0.15386029327005746,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.9453473543978153,
-      "sentence_nr": 6
+      "score": 0.43911482594829104,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
-      "sentence_nr": 6
+      "score": 0.10070927557742705,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
-      "sentence_nr": 6
+      "score": 0.43718220262892105,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.19464521962073492,
-      "sentence_nr": 6
+      "score": 0.0,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_from",
-      "metric": "chrf",
-      "score": 0.5838790966762375,
-      "sentence_nr": 6
-    },
-    {
-      "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_from",
-      "metric": "bleu",
-      "score": 0.3390387389794623,
-      "sentence_nr": 6
-    },
-    {
-      "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_from",
-      "metric": "chrf",
-      "score": 0.6170420596680538,
-      "sentence_nr": 6
-    },
-    {
-      "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_from",
-      "metric": "bleu",
-      "score": 0.3142665434344143,
-      "sentence_nr": 6
-    },
-    {
-      "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_from",
-      "metric": "chrf",
-      "score": 0.6466526067220029,
-      "sentence_nr": 6
-    },
-    {
-      "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_from",
-      "metric": "bleu",
-      "score": 0.4216890913810254,
-      "sentence_nr": 6
-    },
-    {
-      "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6885217194158456,
-      "sentence_nr": 6
+      "score": 0.3370100422576744,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.5014756677893482,
-      "sentence_nr": 6
+      "score": 0.0,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7958858211784339,
-      "sentence_nr": 6
+      "score": 0.1946966569103724,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.6255340042200862,
-      "sentence_nr": 6
+      "score": 0.0772718393063023,
+      "sentence_nr": 1
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.8724783049357475,
-      "sentence_nr": 6
+      "score": 0.4203683137304257,
+      "sentence_nr": 1
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "en",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.1582866049832572,
-      "sentence_nr": 7
+      "score": 0.38870674200492367,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "en",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.34487142413575794,
-      "sentence_nr": 7
+      "score": 0.6484380084879691,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "en",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.17905278399134197,
-      "sentence_nr": 7
+      "score": 0.3961285597009415,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "en",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.37257295447029826,
-      "sentence_nr": 7
+      "score": 0.6148751441350505,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "en",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.15521606028436608,
-      "sentence_nr": 7
+      "score": 0.4923751299732868,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "en",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.37645329404497957,
-      "sentence_nr": 7
+      "score": 0.6853756490381199,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "zh",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.08860973467526746,
-      "sentence_nr": 7
+      "score": 0.11133996756497437,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "zh",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3178004360288637,
-      "sentence_nr": 7
+      "score": 0.4410280353998367,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "zh",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.1418524086391329,
-      "sentence_nr": 7
+      "score": 0.17374951565433233,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "zh",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.38295770773758747,
-      "sentence_nr": 7
+      "score": 0.45325597884524305,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "zh",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.15268019045355535,
-      "sentence_nr": 7
+      "score": 0.17743299460161885,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "zh",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.41028757620299977,
-      "sentence_nr": 7
+      "score": 0.43071271897416463,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "hi",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.19074380068002203,
-      "sentence_nr": 7
+      "score": 0.26459538953931094,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "hi",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.40566585096277824,
-      "sentence_nr": 7
+      "score": 0.5272178908335121,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "hi",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.17382347640129553,
-      "sentence_nr": 7
+      "score": 0.26801022984888695,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "hi",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.4061580777885601,
-      "sentence_nr": 7
+      "score": 0.5654883864995515,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "hi",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.13868172938464635,
-      "sentence_nr": 7
+      "score": 0.21665407194210906,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "hi",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3094469764260441,
-      "sentence_nr": 7
+      "score": 0.4344921442639243,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.13410301071131794,
-      "sentence_nr": 8
+      "score": 0.3563758622144919,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3942932268034351,
-      "sentence_nr": 8
+      "score": 0.6037023613177924,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.598931508663349,
-      "sentence_nr": 8
+      "score": 0.3574583793293068,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7353063745802827,
-      "sentence_nr": 8
+      "score": 0.5924115119819969,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.30677064886592076,
-      "sentence_nr": 8
+      "score": 0.37994652561206577,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5308555945242818,
-      "sentence_nr": 8
+      "score": 0.6464467277069994,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.17328174803055044,
-      "sentence_nr": 8
+      "score": 0.2158914621804855,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3178268797869574,
-      "sentence_nr": 8
+      "score": 0.5448184155666022,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.4262221594184117,
-      "sentence_nr": 8
+      "score": 0.22292726306270316,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5886657414856064,
-      "sentence_nr": 8
+      "score": 0.5653789747970112,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.286608441075188,
-      "sentence_nr": 8
+      "score": 0.09362261118571368,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.4579283646292802,
-      "sentence_nr": 8
+      "score": 0.3452056942265759,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.3315037521841549,
-      "sentence_nr": 8
+      "score": 0.18031307339768174,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.468197879470805,
-      "sentence_nr": 8
+      "score": 0.522164454804456,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.43186481103649477,
-      "sentence_nr": 8
+      "score": 0.21403222128228389,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5792139686527714,
-      "sentence_nr": 8
+      "score": 0.563121432204311,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.30890092021323623,
-      "sentence_nr": 8
+      "score": 0.18917620656425485,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "ur",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5553909583113487,
-      "sentence_nr": 8
+      "score": 0.4346170232980484,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.20972571494011877,
-      "sentence_nr": 9
+      "score": 0.2999092588227898,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.395894071208527,
-      "sentence_nr": 9
+      "score": 0.5505916495384416,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.18559542135951204,
-      "sentence_nr": 9
+      "score": 0.4054983797456263,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3804842882867387,
-      "sentence_nr": 9
+      "score": 0.6264774230839022,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.420450507904553,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.6503146347305717,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "bleu",
       "score": 0.0,
-      "sentence_nr": 9
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4263684749347053,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.20051119758906127,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5334791309401924,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.24894072982768842,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.35369375385786006,
-      "sentence_nr": 9
+      "score": 0.5212235893093335,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.12475846123062707,
-      "sentence_nr": 9
+      "score": 0.2562849004088193,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.27823340731817514,
-      "sentence_nr": 9
+      "score": 0.5767019342009202,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.15122189206102096,
-      "sentence_nr": 9
+      "score": 0.3535002370419364,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.26750110507308866,
-      "sentence_nr": 9
+      "score": 0.5959879218348465,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.10759927692349745,
-      "sentence_nr": 9
+      "score": 0.393613605227227,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "pt",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.21065794536310511,
-      "sentence_nr": 9
+      "score": 0.6492198447661237,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.30327872414714485,
-      "sentence_nr": 9
+      "score": 0.2465888500427759,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.49804213541579834,
-      "sentence_nr": 9
+      "score": 0.5221084445696768,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.21685485833927476,
-      "sentence_nr": 9
+      "score": 0.35983766090218355,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3714219747170047,
-      "sentence_nr": 9
+      "score": 0.5862251404739759,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "bleu",
-      "score": 0.12274092982883021,
-      "sentence_nr": 9
+      "score": 0.21147734744561483,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "pa",
       "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3385513651938691,
-      "sentence_nr": 9
+      "score": 0.41020178654369294,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "en",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.5745954681260859,
-      "sentence_nr": 0
+      "score": 0.38249626297768063,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "en",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7920051188244848,
-      "sentence_nr": 0
+      "score": 0.40976234193505356,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "en",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.6358921902612438,
-      "sentence_nr": 0
+      "score": 0.7281051247089317,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "en",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.8041899227402122,
-      "sentence_nr": 0
+      "score": 0.7882997401328445,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "en",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.6299285159340671,
-      "sentence_nr": 0
+      "score": 0.5806197937310393,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "en",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7993134129243716,
-      "sentence_nr": 0
+      "score": 0.7346706700987636,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.3572514590810421,
-      "sentence_nr": 0
+      "score": 0.08635800047213174,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.40312319760122833,
-      "sentence_nr": 0
+      "score": 0.218109371254876,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.35059076445515835,
-      "sentence_nr": 0
+      "score": 0.25552199116069907,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.40219803477483124,
-      "sentence_nr": 0
+      "score": 0.3799133205289109,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.41316127706749806,
-      "sentence_nr": 0
+      "score": 0.23386786214190372,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.4430321339435623,
-      "sentence_nr": 0
+      "score": 0.3682311523733465,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.9878765474230741,
-      "sentence_nr": 0
+      "score": 0.5642761727828352,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.9958930217841712,
-      "sentence_nr": 0
+      "score": 0.6181373706707737,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.8780634320789833,
-      "sentence_nr": 0
+      "score": 0.4093301993048525,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.926946700115022,
-      "sentence_nr": 0
+      "score": 0.512762518189388,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.7964573357809173,
-      "sentence_nr": 0
+      "score": 0.6244631487487835,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.8458636471716781,
-      "sentence_nr": 0
+      "score": 0.6931369519059803,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.28822910320599077,
-      "sentence_nr": 1
+      "score": 0.2615858282579583,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6087031937056202,
-      "sentence_nr": 1
+      "score": 0.35447530946908884,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.3880515884750121,
-      "sentence_nr": 1
+      "score": 0.4577275269488853,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6587916715823183,
-      "sentence_nr": 1
+      "score": 0.6747054474171109,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.5142726846179982,
-      "sentence_nr": 1
+      "score": 0.25383339228798274,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7344716263345912,
-      "sentence_nr": 1
+      "score": 0.45896379476820603,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.4094748015187699,
-      "sentence_nr": 1
+      "score": 0.15138514598766048,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.4288513205758089,
-      "sentence_nr": 1
+      "score": 0.3237497764315872,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.4487746167679644,
-      "sentence_nr": 1
+      "score": 0.27668736912821895,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.4476730201191672,
-      "sentence_nr": 1
+      "score": 0.4414406760568898,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.2836623400057614,
-      "sentence_nr": 1
+      "score": 0.17200767571780612,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.29147337237183046,
-      "sentence_nr": 1
+      "score": 0.3723150838362789,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.16950698451288215,
-      "sentence_nr": 1
+      "score": 0.15604242268653643,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.48668984177868246,
-      "sentence_nr": 1
+      "score": 0.2255928425212252,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.2113054108348111,
-      "sentence_nr": 1
+      "score": 0.4751132438608344,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.44238229987470284,
-      "sentence_nr": 1
+      "score": 0.6159319815107203,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.26207903587847736,
-      "sentence_nr": 1
+      "score": 0.1477219991186121,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.50073123223194,
-      "sentence_nr": 1
+      "score": 0.28685201698226354,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.6837528314895732,
-      "sentence_nr": 2
+      "score": 0.24728515687112834,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7968789890147058,
-      "sentence_nr": 2
+      "score": 0.3088155734423375,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.7017829861193574,
-      "sentence_nr": 2
+      "score": 0.7281051247089317,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7743327021667388,
-      "sentence_nr": 2
+      "score": 0.78479833664205,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.6961795371760597,
-      "sentence_nr": 2
+      "score": 0.3254455687469726,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7859480663394858,
-      "sentence_nr": 2
+      "score": 0.4474512036484817,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.2360941227140328,
-      "sentence_nr": 2
+      "score": 0.353203510510529,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.35939098278145853,
-      "sentence_nr": 2
+      "score": 0.4910213297498164,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.14118350058219528,
-      "sentence_nr": 2
+      "score": 0.4815092081725061,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.20431837779877604,
-      "sentence_nr": 2
+      "score": 0.5820265218174012,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.1811004938014804,
-      "sentence_nr": 2
+      "score": 0.23887527917609022,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.2649993136544717,
-      "sentence_nr": 2
+      "score": 0.4120359948636439,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.23649053182388327,
-      "sentence_nr": 2
+      "score": 0.23660362391696813,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.4127382174759535,
-      "sentence_nr": 2
+      "score": 0.34152697838249696,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.20721924345714232,
-      "sentence_nr": 2
+      "score": 0.580451128369423,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.36475932190367044,
-      "sentence_nr": 2
+      "score": 0.7246473808162345,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.11386607947762988,
-      "sentence_nr": 2
+      "score": 0.580451128369423,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.33564583347921473,
-      "sentence_nr": 2
+      "score": 0.728208634600343,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.8107492451395732,
+      "score": 1.0,
       "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.900032747778274,
+      "score": 1.0,
       "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.8761560783209453,
+      "score": 0.8482942955247808,
       "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.9484564543183253,
+      "score": 0.9256238040654331,
       "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.7505336182671021,
+      "score": 1.0,
       "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.8401910628269498,
+      "score": 1.0,
       "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.13725861056573663,
-      "sentence_nr": 3
+      "score": 0.8003203203844999,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.11147384852362276,
-      "sentence_nr": 3
+      "score": 0.9453478043428296,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.13453927150397377,
-      "sentence_nr": 3
+      "score": 1.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.10522974272748564,
-      "sentence_nr": 3
+      "score": 1.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.22055493694673897,
-      "sentence_nr": 3
+      "score": 1.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3931965048763613,
-      "sentence_nr": 3
+      "score": 1.0,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.839587623092576,
-      "sentence_nr": 3
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.9096086668952811,
-      "sentence_nr": 3
+      "score": 0.629039349740581,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.6885326214539055,
-      "sentence_nr": 3
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.8229812189228393,
-      "sentence_nr": 3
+      "score": 0.629039349740581,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.839587623092576,
-      "sentence_nr": 3
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.9096086668952811,
-      "sentence_nr": 3
+      "score": 0.629039349740581,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.8003203203844999,
+      "score": 0.0,
       "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.9453478043428296,
+      "score": 0.49546288984677567,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
+      "score": 0.0,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
+      "score": 0.43795381992037963,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
+      "score": 0.0,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
+      "score": 0.5881561248602009,
       "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.7215691881328408,
+      "score": 0.0,
       "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.735100789804592,
+      "score": 0.40854152133685306,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.4765874091118851,
+      "score": 0.0,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.45911557772276623,
+      "score": 0.41213231348812146,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.45022125383821326,
+      "score": 0.0,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.46874267375238576,
+      "score": 0.40435987083533204,
       "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.2028736642487601,
+      "score": 0.0,
       "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.43458947791319813,
+      "score": 0.39618802899930716,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "bleu",
       "score": 0.0,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3013901676230198,
+      "score": 0.39618802899930716,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "bleu",
       "score": 0.0,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.38106012955734714,
+      "score": 0.39858613265631837,
       "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
-      "sentence_nr": 5
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
-      "sentence_nr": 5
+      "score": 0.5948724602646328,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.2887138086538547,
-      "sentence_nr": 5
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6342291345998248,
-      "sentence_nr": 5
+      "score": 0.5042211795038526,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
-      "sentence_nr": 5
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
-      "sentence_nr": 5
+      "score": 0.4425973012069069,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.20748131961458333,
-      "sentence_nr": 5
+      "score": 0.27447938256311044,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.2716205232346228,
-      "sentence_nr": 5
+      "score": 0.615291848344044,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.37589902061551017,
-      "sentence_nr": 5
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.42554151277542873,
-      "sentence_nr": 5
+      "score": 0.5516607622642397,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.20748131961458333,
-      "sentence_nr": 5
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.2716205232346228,
-      "sentence_nr": 5
+      "score": 0.47160616105623426,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.4135171000263379,
-      "sentence_nr": 5
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7050151549073953,
-      "sentence_nr": 5
+      "score": 0.33762297226992255,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.42988105429544615,
-      "sentence_nr": 5
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7577244658187771,
-      "sentence_nr": 5
+      "score": 0.4576529535952892,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.5366411241731205,
-      "sentence_nr": 5
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.825566494253596,
-      "sentence_nr": 5
+      "score": 0.5309982646782259,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
-      "sentence_nr": 6
+      "score": 0.2658483576665877,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
-      "sentence_nr": 6
+      "score": 0.6410540990527072,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.7511573912724299,
-      "sentence_nr": 6
+      "score": 0.24601372576927547,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.9453473543978153,
-      "sentence_nr": 6
+      "score": 0.6374693500772332,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 1.0,
-      "sentence_nr": 6
+      "score": 0.1892240568795935,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "pt",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 1.0,
-      "sentence_nr": 6
+      "score": 0.6151179643430991,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.3450219162509876,
-      "sentence_nr": 6
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.3993348853061597,
-      "sentence_nr": 6
+      "score": 0.6562641136790542,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.5561195823338172,
-      "sentence_nr": 6
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5362935676066722,
-      "sentence_nr": 6
+      "score": 0.46426595961938383,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.5803515898273521,
-      "sentence_nr": 6
+      "score": 0.0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "pa",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5422220468910552,
-      "sentence_nr": 6
+      "score": 0.41238100267720657,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.2562150245540302,
-      "sentence_nr": 6
+      "score": 1.0,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.47046477830594896,
-      "sentence_nr": 6
+      "score": 1.0,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.33438299066966715,
-      "sentence_nr": 6
+      "score": 0.2887138086538547,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5409759573191787,
-      "sentence_nr": 6
+      "score": 0.6342291345998248,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.37854068916316835,
-      "sentence_nr": 6
+      "score": 1.0,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "en",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5743796566387722,
-      "sentence_nr": 6
+      "score": 1.0,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.44711013370113256,
-      "sentence_nr": 7
+      "score": 0.47410002229034043,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7319347493436125,
-      "sentence_nr": 7
+      "score": 0.7538467008030766,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.49023502313124495,
-      "sentence_nr": 7
+      "score": 0.5087473540251254,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.7638414724136195,
-      "sentence_nr": 7
+      "score": 0.7647955332172516,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.4424906782646928,
-      "sentence_nr": 7
+      "score": 0.5087473540251254,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "zh",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.705507971295129,
-      "sentence_nr": 7
+      "score": 0.7647955332172516,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.21326369102393236,
-      "sentence_nr": 7
-    },
+      "score": 0.0,
+      "sentence_nr": 5
+    },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.24781828193168487,
-      "sentence_nr": 7
+      "score": 0.2828367156737383,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.07860105393900486,
-      "sentence_nr": 7
+      "score": 0.48181149445310956,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.09678377693633947,
-      "sentence_nr": 7
+      "score": 0.7675828789334244,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.11601141307045003,
-      "sentence_nr": 7
+      "score": 0.5091224918749461,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "hi",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.21671187566850864,
-      "sentence_nr": 7
+      "score": 0.7829685247145245,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.2577716972449781,
-      "sentence_nr": 7
+      "score": 0.44897710722021167,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5171901208397282,
-      "sentence_nr": 7
+      "score": 0.6862249089515978,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.35015224715252113,
-      "sentence_nr": 7
+      "score": 0.404727200247809,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5701648579139658,
-      "sentence_nr": 7
+      "score": 0.6681898017773897,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.3349252032650068,
-      "sentence_nr": 7
+      "score": 0.40276720463657734,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "es",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5908087431574293,
-      "sentence_nr": 7
+      "score": 0.6529271690805427,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.22816849039973935,
-      "sentence_nr": 8
+      "score": 0.2704091953828695,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5295534280606148,
-      "sentence_nr": 8
+      "score": 0.6207272323003366,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.348007986647201,
-      "sentence_nr": 8
+      "score": 0.5379348324975908,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6148736550683231,
-      "sentence_nr": 8
+      "score": 0.7703766110349561,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.31222258402876674,
-      "sentence_nr": 8
+      "score": 0.30188353873287377,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "ar",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5549937870516303,
-      "sentence_nr": 8
+      "score": 0.6086565367747951,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.0,
-      "sentence_nr": 8
+      "score": 0.48181149445310956,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.10721126066665879,
-      "sentence_nr": 8
+      "score": 0.7675828789334244,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.23683075175361493,
-      "sentence_nr": 8
+      "score": 0.5091224918749461,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.2631328190836655,
-      "sentence_nr": 8
+      "score": 0.7829685247145245,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.0,
-      "sentence_nr": 8
+      "score": 0.6026286934891149,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "translation_to",
+      "bcp_47": "ur",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.16455392433653304,
-      "sentence_nr": 8
+      "score": 0.8025775976044891,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.36033217429111203,
-      "sentence_nr": 8
+      "score": 0.4596980088392874,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5707860320039717,
-      "sentence_nr": 8
+      "score": 0.713787745993602,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.461597801606675,
-      "sentence_nr": 8
+      "score": 0.5896613549548209,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6280777654467244,
-      "sentence_nr": 8
+      "score": 0.7528914749586836,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.4224298950114519,
-      "sentence_nr": 8
+      "score": 0.4596980088392874,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "translation_to",
+      "bcp_47": "fr",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.60823085524287,
-      "sentence_nr": 8
+      "score": 0.713787745993602,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.43103580001357805,
-      "sentence_nr": 9
+      "score": 0.0,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.6690742226623104,
-      "sentence_nr": 9
+      "score": 0.24706467963183681,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.6031612036218008,
-      "sentence_nr": 9
+      "score": 0.4801289744823913,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.736286703381354,
-      "sentence_nr": 9
+      "score": 0.6766690087429765,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.39432344823662835,
-      "sentence_nr": 9
+      "score": 0.3272712268138726,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "translation_to",
+      "bcp_47": "bn",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5943452555220106,
-      "sentence_nr": 9
+      "score": 0.6272846474183881,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.3665134361137304,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.6118771029352303,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.28489318277723963,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5764325110247531,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.2981792160679168,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5788026000794341,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.2799331151961311,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.6471892368478446,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.8142499721936278,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.7012294787544179,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.8478115719875968,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.7511573912724299,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.9453473543978153,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.10772332006118607,
-      "sentence_nr": 9
+      "score": 0.19464521962073492,
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.17652714369664665,
-      "sentence_nr": 9
+      "score": 0.5838790966762375,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.11254397891886614,
-      "sentence_nr": 9
+      "score": 0.3390387389794623,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.20623288988983426,
-      "sentence_nr": 9
+      "score": 0.6170420596680538,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.1100081929352474,
-      "sentence_nr": 9
+      "score": 0.3142665434344143,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "zh",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.18967061672400035,
-      "sentence_nr": 9
+      "score": 0.6466526067220029,
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.27190910124573536,
-      "sentence_nr": 9
+      "score": 0.4216890913810254,
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.5173567851798608,
-      "sentence_nr": 9
+      "score": 0.6885217194158456,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.24493390281390082,
-      "sentence_nr": 9
+      "score": 0.5014756677893482,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.48113625107113883,
-      "sentence_nr": 9
+      "score": 0.7958858211784339,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "bleu",
-      "score": 0.19476681308252697,
-      "sentence_nr": 9
+      "score": 0.6255340042200862,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "hi",
-      "task": "translation_to",
+      "task": "translation_from",
       "metric": "chrf",
-      "score": 0.42030407727741037,
-      "sentence_nr": 9
+      "score": 0.8724783049357475,
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "classification",
-      "metric": "accuracy",
-      "score": 1,
-      "sentence_nr": 0
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.32365795029773287,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.6509517796070665,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "classification",
-      "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 0
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.48994561421713123,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.7411155087367244,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "classification",
-      "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 0
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.48994561421713123,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.7411155087367244,
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.44787223195695314,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.7968980206907678,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.5971070986250356,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.8874294965619517,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.5971070986250356,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.8874294965619517,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.5088645484558708,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.6991726442472661,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.4101479464529936,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.7041976254287654,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.4547900039222725,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.6541971428810075,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.23198210427894825,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.630711601223299,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.4284945090100314,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.7164026439677106,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.23198210427894825,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.630711601223299,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.48994561421713123,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.8020845125558708,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.595092211343687,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.7971172820981081,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.4831233610237384,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.7122562458056777,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.44787223195695314,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.7968980206907678,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.5971070986250356,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.8874294965619517,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.32263864160302524,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.6824395076981005,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.5383680940297331,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.786096406361039,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.5383680940297331,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.786096406361039,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.38305978177479755,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.6061131723054572,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.1582866049832572,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.34487142413575794,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.17905278399134197,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.37257295447029826,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.15521606028436608,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.37645329404497957,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.08860973467526746,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3178004360288637,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.1418524086391329,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.38295770773758747,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.15268019045355535,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.41028757620299977,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.19074380068002203,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.40566585096277824,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.17382347640129553,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4061580777885601,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.13868172938464635,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3094469764260441,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.22381487678101888,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5249370100068887,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.09431297723472011,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3616856339096348,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.11091252683001185,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.26607634610445896,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.10666682719585797,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.33462901494141756,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.14557808399334188,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.36598346755702993,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.18154954789336694,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4557483776072868,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.20198948917565754,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.34858221035657466,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.16780109158842918,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3968694014697679,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.1381751568911733,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3121557499162649,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.22669486951066523,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4484451941575473,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.11697642623186386,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.37117753637984835,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.20065115069964384,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4084885616013531,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.09916009482330297,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3032928217006101,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.13805615693046389,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.40787998733941394,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.12291219097556666,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3448002180666873,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.09478705591775652,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.33293232395887284,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.19148282873929853,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4707949702068854,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.20608572305725564,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4704943905570542,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.14057105892389254,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3028381427383384,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.12157241570357182,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4080990097991491,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.085416483900781,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.2825804066750608,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.13410301071131794,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3942932268034351,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.598931508663349,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.7353063745802827,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.30677064886592076,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5308555945242818,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.17328174803055044,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3178268797869574,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.4262221594184117,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5886657414856064,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.286608441075188,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4579283646292802,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.3315037521841549,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.468197879470805,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.43186481103649477,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5792139686527714,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.30890092021323623,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5553909583113487,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.22218130727359342,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.39929356245904674,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.46092611919700416,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.6365915338629015,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.1690979933029136,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3751861276375209,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.26538706048179084,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4982627378595717,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.5234484809182233,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.6658297773613274,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.2840563956846642,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5110250591004448,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.15161074985415177,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3796830006266126,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.3815250264738168,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.6516314751979607,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.19920413481788912,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.42537796926163113,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.18679710353734788,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3876457319870774,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.40003810431098236,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5899097408105687,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.20401796878756984,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.43317630453631556,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.21812881407613688,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3598346059855135,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.38047531731529327,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.49485723102957346,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.2044887070217883,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.38471585132587544,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.24586918158076287,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4658595745396681,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.4489235959690452,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5934678825154104,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.2980504190448601,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.5101268920225042,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.46832763312452297,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.8176110134774669,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.9436043261706615,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.9880191679951993,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.20972571494011877,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.395894071208527,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.18559542135951204,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3804842882867387,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.35369375385786006,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.12475846123062707,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.27823340731817514,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.15122189206102096,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.26750110507308866,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.10759927692349745,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.21065794536310511,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.30327872414714485,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.49804213541579834,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.21685485833927476,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3714219747170047,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.12274092982883021,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3385513651938691,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.12787395553510186,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.33752742535974617,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.112289032173749,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.17726100052085036,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.31017716089889963,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.39962545473912425,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3710595252626966,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.21688283061839067,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.41775824162589076,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.14326513489612383,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4034278533385552,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.12666372160329223,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.2650373529479294,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.12162779391619735,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3228288840559658,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.18294117097472648,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.4383387744769579,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.203264842568494,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.2922087191170089,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.18237599479708327,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3740403511567824,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.20298407172594946,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.427376330935813,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.1740044679403827,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.36375152376157177,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.12876689524369925,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3253153379449275,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.15626231814206226,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.2918712789926548,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.322788951728102,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.40263021320001785,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.119159749312327,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.21297942664093145,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.21397099133614067,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3568171392601981,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.16925466459550803,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.35912398848424326,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "bleu",
+      "score": 0.2036348471340078,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_from",
+      "metric": "chrf",
+      "score": 0.3472831655579266,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5745954681260859,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7920051188244848,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6358921902612438,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8041899227402122,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6299285159340671,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7993134129243716,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3572514590810421,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.40312319760122833,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.35059076445515835,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.40219803477483124,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.41316127706749806,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4430321339435623,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.9878765474230741,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.9958930217841712,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.8780634320789833,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.926946700115022,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.7964573357809173,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8458636471716781,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.35601247064914876,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6528728847159075,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.40673971192998765,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6897190926100627,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3707525915417785,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6481906761834414,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5724622291345857,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6818279156433621,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.703373719677874,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7784050705257474,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.469958733898233,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5843756060033074,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2676232320051144,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5440246804235981,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3489926819498492,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5715668842319502,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2786169604662155,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5267252236203236,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3765213224289163,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6469521424555786,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3410244689880313,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5816669416914216,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4234343012313773,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6625289905598352,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.44219732271776674,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6193429426274062,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4324680011853555,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5877600878871951,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4493940083619696,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6230960824462234,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6638859619095425,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7874224590682172,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6947677373756656,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7941300666655116,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6412098671661826,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7665040244283648,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4845227999608418,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5968050469845498,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3861375213265022,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5122109329134508,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.32539921259497445,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5133457276293165,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.28822910320599077,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6087031937056202,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3880515884750121,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6587916715823183,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5142726846179982,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7344716263345912,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4094748015187699,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4288513205758089,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4487746167679644,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4476730201191672,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2836623400057614,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.29147337237183046,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.16950698451288215,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.48668984177868246,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2113054108348111,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.44238229987470284,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.26207903587847736,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.50073123223194,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4527112325797497,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6708989870027865,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4556160153884204,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6661994452325181,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3216756020053242,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6141241026166391,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3977038258772401,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6202897864314184,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.37570809340937233,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6339141734561076,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.559332422592187,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.733291190094771,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3167585643537871,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5076869840147092,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3446592076818278,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5819912583909785,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.23270938096152352,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4490269267329941,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.44114781827798216,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6241365710582877,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.529527758323629,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6540432510655854,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.49704232910799745,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6453248294274054,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3542266508664836,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5643413028542406,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3479698393875884,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5760833125751785,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.24373253714463095,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.49482039214573803,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3995439803178399,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6021193793256325,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2988697040013311,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5442522660489195,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.500703635659656,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6501904887399698,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2993081268625724,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.47777429598730525,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.43330223254789785,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5564499529933307,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.1466607445607986,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.36552963821230766,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6837528314895732,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7968789890147058,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.7017829861193574,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7743327021667388,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6961795371760597,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7859480663394858,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2360941227140328,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.35939098278145853,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.14118350058219528,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.20431837779877604,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.1811004938014804,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.2649993136544717,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.23649053182388327,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4127382174759535,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.20721924345714232,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.36475932190367044,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.11386607947762988,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.33564583347921473,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.570135897056151,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6801332690579707,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.46442643702863534,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5519480629125156,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6268941789647348,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6958291103494518,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.21305368975019265,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4371748197696026,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.22837680015088951,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.44164180234500505,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4151474543103342,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.638952468710771,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.20876900081884944,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3981381071356935,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.11634129390828839,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.31530902302000635,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.19544795798162903,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3835451743665027,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.37917766663411384,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5365794450039074,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.359355103997122,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5589602235417395,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4267520229161,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5518115366540288,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3221305290185444,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4866081657424789,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.35551034193127495,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5627284645723449,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5039752490702457,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.613669501327356,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6260375038358343,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7803415401430737,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5088535943352446,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.625202596789752,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.562048819850726,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7192054483864224,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3146726146646545,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4709531555683,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3941975148525721,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5191046479503385,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.262633940062176,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.41923206553744197,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.8107492451395732,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.900032747778274,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.8761560783209453,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.9484564543183253,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.7505336182671021,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8401910628269498,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.13725861056573663,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.11147384852362276,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.13453927150397377,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.10522974272748564,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.22055493694673897,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3931965048763613,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.839587623092576,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.9096086668952811,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6885326214539055,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8229812189228393,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.839587623092576,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.9096086668952811,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.412295470431275,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.705800771033924,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3480442076026084,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6142483232997242,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2861853478258715,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6401604432917332,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.14172292406325543,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4762857001428092,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.17401517708317762,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.45006261596496794,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.151240443751577,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4224869587588239,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.21850594525107195,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4049269026117245,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2028736642487601,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3614856639698008,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2255489037266197,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3954925749722234,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.2466674257522263,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2929807168354841,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5975595069845072,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6689604664235209,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8010329764520807,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4678134833959513,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5051480556620123,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4113125177363443,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.42808075762838727,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.32685141385924577,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3758692873615971,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.27341185048222727,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6411651849711889,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6237774736059616,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8500131524897436,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.573764722928549,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.817979859532479,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.8482942955247808,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.9256238040654331,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.8003203203844999,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.9453478043428296,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.7215691881328408,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.735100789804592,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4765874091118851,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.45911557772276623,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.45022125383821326,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.46874267375238576,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2028736642487601,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.43458947791319813,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3013901676230198,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.38106012955734714,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3820562306791339,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5488313413379253,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3820562306791339,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5488313413379253,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.38091370416670794,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.540550443602966,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.23189835231884592,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.44157797833899437,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2294068720558097,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.40458364050078693,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.32588643749980295,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.31430120091187586,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5013155459452984,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.45237912327122276,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6299071573751139,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.23736810439041953,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4004852416401387,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3722001929300059,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5252698638532942,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5605065818946205,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.586853267829013,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.36300296341860155,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5816676674074003,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.1258907882951215,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.34143648068854054,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.28765408533715414,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4046608868073569,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.1819722649161304,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.44254730215235283,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.28295596283263513,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6067794553589253,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.13305199541830684,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.43244987270004115,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3377385620641691,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5256128450453542,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.1175771442804648,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3103572690939351,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.12546912767038895,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.2651343523961406,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.12407216162020399,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.2664864612493293,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2887138086538547,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6342291345998248,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.20748131961458333,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.2716205232346228,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.37589902061551017,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.42554151277542873,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.20748131961458333,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.2716205232346228,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4135171000263379,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7050151549073953,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.42988105429544615,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7577244658187771,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5366411241731205,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.825566494253596,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6976333495952621,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8331572107884448,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4165530720734658,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7027805129995731,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4027788021844849,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6872835607174038,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.498704623570665,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6478746389895599,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.41307323705325416,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5785653391533346,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5248587176134882,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6664855309004869,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.35210829264331733,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5239651686730163,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.17729842264695017,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4103582047611184,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.34895836374229405,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4767378358574124,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.833078701050083,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.9482515348146272,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.8958039312312598,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.9382091007325469,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5805399561362194,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4810464260105228,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4500531895417844,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.43027065541050147,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4933292241270431,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5225247297523148,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5064127215831256,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6516332048338376,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5212982931053122,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.35319015092357736,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5822934956325967,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4592978565863154,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3855522725905196,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.587260566914102,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4426623526629488,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6368371029698285,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.7511573912724299,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.9453473543978153,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3450219162509876,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3993348853061597,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5561195823338172,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5362935676066722,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5803515898273521,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5422220468910552,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2562150245540302,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.47046477830594896,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.33438299066966715,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5409759573191787,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.37854068916316835,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5743796566387722,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5521710658453207,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7317828775912516,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.45506803308128024,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6477506541284608,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.592313615748771,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7382416555842614,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4018202851356865,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6003256951549871,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4018202851356865,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6003256951549871,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.580451128369423,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7542976177437886,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.18207052811092134,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4504432021668592,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.37717457428685847,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5554130492458337,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.31598923484911084,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6425503166524515,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8078891929749037,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.7629273292796576,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8510385544954956,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.8725129388059689,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.9495292423959529,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2147607499133801,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3976144917079093,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.49349163706233623,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.694445271037971,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3037643089519314,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5183662698462751,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.8253498772794055,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8529564805429163,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.7944837206494969,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8784531740275225,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.23887527917609022,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5924993690004501,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5828833474188783,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7908226509294533,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.38694317759010316,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5953878513137957,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.44711013370113256,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7319347493436125,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.49023502313124495,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7638414724136195,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4424906782646928,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.705507971295129,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.21326369102393236,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.24781828193168487,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.07860105393900486,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.09678377693633947,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.11601141307045003,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.21671187566850864,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2577716972449781,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5171901208397282,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.35015224715252113,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5701648579139658,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3349252032650068,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5908087431574293,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3274016883618531,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5570399656004248,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3449058130015412,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5365619830343804,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.29688845677442144,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5494319015457763,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.207314191412716,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4360555836773355,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.25376032254696296,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5334329403985332,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.20039141607873007,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.36123312088832493,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3235473265529593,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5441122251341168,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.32707695373369694,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5166643606783462,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3069937936246452,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5024648105961349,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.26513488970168847,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6029932145447834,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3927237741677927,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7451438087039315,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5570357635362685,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8116469942298856,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3860973950960897,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6271680934322363,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3826576187198625,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6071841372061269,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3447241447679157,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5531085140985558,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.27710310401156996,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5216248191624099,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.31128635710849173,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6304411194127884,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.13308561809919006,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5312476702183977,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.20475739007221866,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3934874462686164,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2879556779114461,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4554184077174173,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.09578921953028982,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.40472887922389433,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.22816849039973935,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5295534280606148,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.348007986647201,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6148736550683231,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.31222258402876674,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5549937870516303,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.10721126066665879,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.23683075175361493,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.2631328190836655,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.16455392433653304,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.36033217429111203,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5707860320039717,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.461597801606675,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6280777654467244,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4224298950114519,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.60823085524287,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4141871474340027,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.36769040719718776,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4064141882459388,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3096036988813059,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5894510883198948,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3286711939680359,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5944310794747374,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.13547277341758465,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4830189619506113,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.23841754841770157,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.34481325534410395,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2822535302220024,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3874773378787974,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.31747697264511426,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.40797778663955364,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.43368945552925614,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6793717376740783,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.46935933364934335,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.773055573548356,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.43368945552925614,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.640995178057518,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.20050320605789015,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4046291070099031,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3160213610127146,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5165614670038283,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.30758744700466467,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4684197705189288,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4161791450287817,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.7054426787013603,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2961516536011624,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.49803924348035766,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3459667618766101,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6170810606402402,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 1.0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 1.0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.46832763312452297,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.8176110134774669,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.9436043261706615,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.9880191679951993,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.43103580001357805,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6690742226623104,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.6031612036218008,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.736286703381354,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.39432344823662835,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5943452555220106,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.10772332006118607,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.17652714369664665,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.11254397891886614,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.20623288988983426,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.1100081929352474,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.18967061672400035,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.27190910124573536,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5173567851798608,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.24493390281390082,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.48113625107113883,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.19476681308252697,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.42030407727741037,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.429512074830509,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6066779955199886,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3459789902390003,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5620330456296532,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.37825713491091884,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5584414289480568,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.09596136927307748,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.40849147213099996,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.1845747513433909,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.44379971518505973,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.18212463619188357,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.469592540371137,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.22972631482860506,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.436102988762466,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.2517176762753373,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.45137344500317134,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.3128384316903283,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.48016279207050283,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.45026965676007474,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6668256174353906,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.43200638115383627,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6892273787708799,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.341195158470265,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6539473951166187,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4710260495003035,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.40475700826319555,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.4349871720911447,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.310186302993101,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.5434540129901786,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.5244380103905697,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6914581279144536,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.4472834999328078,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6457130269652316,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.17207258849758605,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.3052503498954155,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.504154287515855,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.6074467585243234,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "bleu",
+      "score": 0.18771816026273827,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "translation_to",
+      "metric": "chrf",
+      "score": 0.37594160796244835,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 4
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 4
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 5
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 5
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 7
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 7
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 8
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 8
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ur",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pa",
+      "task": "classification",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pa",
       "task": "classification",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 0
+      "score": 0,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 0
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "es",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "es",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "ar",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "fr",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "bn",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 1
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 1
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "en",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "en",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 0,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "en",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "zh",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "zh",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "zh",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "hi",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "hi",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "hi",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "es",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "es",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 0
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "es",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 0
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 0
+      "score": 1,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 0
+      "score": 1,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 0
+      "score": 1,
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "fr",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 1
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "fr",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 1
+      "score": 1,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "fr",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 1
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "bn",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 1
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "bn",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 1
+      "score": 1,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "bn",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 1
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 1
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 1
+      "score": 1,
+      "sentence_nr": 2
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 1
+      "sentence_nr": 2
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "en",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 1,
-      "sentence_nr": 2
+      "score": 0,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "en",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 1,
-      "sentence_nr": 2
+      "score": 0,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "en",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 1,
-      "sentence_nr": 2
+      "score": 0,
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "zh",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 2
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "zh",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 2
+      "score": 1,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "zh",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 2
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "hi",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 2
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "hi",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 2
+      "score": 1,
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "hi",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 2
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "es",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
       "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "es",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
       "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "es",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
       "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
       "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
       "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
       "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "fr",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
       "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "fr",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "fr",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
       "sentence_nr": 3
     },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "bn",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "bn",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 3
+    },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "bn",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
       "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 4
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 4
+      "sentence_nr": 3
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 4
+      "sentence_nr": 3
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "en",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "en",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "en",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
       "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "zh",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "zh",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
       "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "zh",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
       "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "hi",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 5
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "hi",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 5
+      "score": 1,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "hi",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 5
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "es",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 5
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "es",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 5
+      "score": 1,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "es",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 5
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 1,
-      "sentence_nr": 5
+      "score": 0,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 5
+      "score": 1,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 1,
-      "sentence_nr": 5
+      "score": 0,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "fr",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 6
+      "score": 1,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "fr",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 6
+      "score": 1,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "fr",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 6
+      "score": 1,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "bn",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 6
+      "score": 1,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "bn",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 6
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "bn",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 6
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 6
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 6
+      "score": 1,
+      "sentence_nr": 4
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 6
+      "score": 1,
+      "sentence_nr": 4
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "en",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "en",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "en",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "zh",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "zh",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "zh",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 7
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "hi",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "hi",
-      "task": "classification",
-      "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
-    },
-    {
-      "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "classification",
-      "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
-    },
-    {
-      "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "classification",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
-    },
-    {
-      "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "classification",
-      "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 8
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "hi",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "es",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "es",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "es",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 8
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "ar",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "fr",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 9
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "fr",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 9
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
-      "task": "classification",
+      "bcp_47": "fr",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 9
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "bn",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 9
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "bn",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 9
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
-      "task": "classification",
+      "bcp_47": "bn",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 9
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 9
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 9
+      "sentence_nr": 5
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
-      "task": "classification",
+      "bcp_47": "pt",
+      "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 9
+      "score": 1,
+      "sentence_nr": 5
     },
     {
       "model": "openai/gpt-4o-mini",
@@ -11889,7 +21441,7 @@
       "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 0
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
@@ -11897,23 +21449,23 @@
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 0
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "en",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 1,
-      "sentence_nr": 0
+      "score": 0,
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "zh",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 1,
-      "sentence_nr": 0
+      "score": 0,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
@@ -11921,15 +21473,15 @@
       "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 0
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "zh",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 1,
-      "sentence_nr": 0
+      "score": 0,
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
@@ -11937,7 +21489,7 @@
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 0
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
@@ -11945,7 +21497,7 @@
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 0
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
@@ -11953,359 +21505,383 @@
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 0
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 1
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 1
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 1
+      "score": 1,
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "ar",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 1
+      "sentence_nr": 6
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "ar",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 6
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "fr",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "fr",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 1
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "fr",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 1
+      "score": 1,
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 1
+      "score": 1,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 1
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 1
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "pt",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 2
+      "score": 1,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "pt",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 2
+      "score": 1,
+      "sentence_nr": 6
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "pt",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 2
+      "score": 1,
+      "sentence_nr": 6
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "en",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 2
+      "score": 1,
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "en",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 2
+      "score": 1,
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "en",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 2
+      "sentence_nr": 7
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "zh",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 2
+      "score": 1,
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "zh",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 2
+      "score": 1,
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "zh",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 2
+      "sentence_nr": 7
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "hi",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 3
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "hi",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 3
+      "score": 1,
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "hi",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 3
+      "sentence_nr": 7
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 3
+      "score": 1,
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 3
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 1,
-      "sentence_nr": 3
+      "score": 0,
+      "sentence_nr": 7
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "ar",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 3
+      "score": 1,
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "ar",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 3
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "ar",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 3
+      "score": 1,
+      "sentence_nr": 7
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 4
+      "score": 1,
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 4
+      "score": 1,
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "fr",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 4
+      "sentence_nr": 7
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 4
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 4
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 1,
-      "sentence_nr": 4
+      "score": 0,
+      "sentence_nr": 7
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "pt",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 0,
-      "sentence_nr": 4
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "pt",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 4
+      "score": 1,
+      "sentence_nr": 7
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "pt",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 4
+      "sentence_nr": 7
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "en",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 5
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "en",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 5
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "en",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 5
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "zh",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 5
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "zh",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 5
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
       "bcp_47": "zh",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 5
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "openai/gpt-4o-mini",
       "bcp_47": "hi",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 5
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-4-maverick",
       "bcp_47": "hi",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 5
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
@@ -12313,294 +21889,318 @@
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 5
+      "sentence_nr": 8
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 6
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 6
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 6
+      "sentence_nr": 8
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 6
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 6
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "ar",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 6
+      "sentence_nr": 8
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "fr",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 6
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "fr",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 6
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "fr",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 6
+      "sentence_nr": 8
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "pt",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "pt",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "pt",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 8
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "en",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "en",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "en",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 7
+      "score": 1,
+      "sentence_nr": 9
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "zh",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "zh",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 8
+      "score": 1,
+      "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "zh",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 9
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "hi",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "hi",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 8
+      "score": 1,
+      "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "hi",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 9
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
-      "sentence_nr": 8
+      "score": 1,
+      "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "es",
       "task": "mmlu",
       "metric": "accuracy",
       "score": 1,
-      "sentence_nr": 8
+      "sentence_nr": 9
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "en",
+      "bcp_47": "ar",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
       "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "en",
+      "bcp_47": "ar",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
       "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "en",
+      "bcp_47": "ar",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
       "sentence_nr": 9
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "zh",
+      "bcp_47": "fr",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
       "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "zh",
+      "bcp_47": "fr",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
       "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "zh",
+      "bcp_47": "fr",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
       "sentence_nr": 9
     },
     {
       "model": "openai/gpt-4o-mini",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
       "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-4-maverick",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
       "sentence_nr": 9
     },
     {
       "model": "meta-llama/llama-3.3-70b-instruct",
-      "bcp_47": "hi",
+      "bcp_47": "bn",
       "task": "mmlu",
       "metric": "accuracy",
-      "score": 0,
+      "score": 1,
+      "sentence_nr": 9
+    },
+    {
+      "model": "openai/gpt-4o-mini",
+      "bcp_47": "pt",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-4-maverick",
+      "bcp_47": "pt",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
+      "sentence_nr": 9
+    },
+    {
+      "model": "meta-llama/llama-3.3-70b-instruct",
+      "bcp_47": "pt",
+      "task": "mmlu",
+      "metric": "accuracy",
+      "score": 1,
       "sentence_nr": 9
     }
   ]