Hugging Face's logo

Datasets:
MTSAIR
/
MWS-Vision-Bench

Tasks:
Visual Question Answering
Document Question Answering
Modalities:
Image
Text
Document
Formats:
imagefolder
Languages:
Russian
Size:
1K - 10K
Tags:
benchmark
multimodal
ocr
kie
grounding
vlm
Libraries:
Datasets
License:
Dataset card Data Studio Files
xet
Community
1
GeorgyG commited on
Commit
2e134f0
·
verified ·
1 Parent(s): ab13d28

Add leaderboard with Qwen3-VL models

Browse files
Files changed (1) hide show
  1. README.md +12 -7
README.md CHANGED
@@ -3,7 +3,7 @@ pretty_name: MWS Vision Bench
3
  dataset_name: mws-vision-bench
4
  language:
5
  - ru
6
- license: mit
7
  tags:
8
  - benchmark
9
  - multimodal
@@ -114,7 +114,9 @@ Each line in `metadata.jsonl` contains one JSON object:
114
 
115
  ---
116
 
117
- ## Leaderboard
 
 
118
 
119
  | Model | Overall | img→text | img→markdown | Grounding | KIE (JSON) | VQA |
120
  |-------|---------|----------|--------------|-----------|------------|-----|
@@ -122,23 +124,26 @@ Each line in `metadata.jsonl` contains one JSON object:
122
  | **Gemini-2.5-flash** | **0.644** | 0.796 | 0.683 | 0.067 | 0.841 | 0.833 |
123
  | **gpt-4.1-mini** | **0.643** | 0.866 | 0.724 | 0.091 | 0.750 | 0.782 |
124
  | **Claude-4.5-Sonnet** | **0.639** | 0.723 | 0.676 | 0.377 | 0.728 | 0.692 |
125
- | **Cotype VL (32B 8 bit)** | **0.639** | 0.797 | 0.756 | 0.262 | 0.694 | 0.685 |
126
- | gpt-5-mini | 0.632 | 0.797 | 0.678 | 0.126 | 0.784 | 0.776 |
127
  | Qwen2.5-VL-72B | 0.631 | 0.848 | 0.712 | 0.220 | 0.644 | 0.732 |
128
  | gpt-5-mini (responses) | 0.594 | 0.743 | 0.567 | 0.118 | 0.811 | 0.731 |
129
- | Qwen3-VL-8B-Instruct | 0.592 | 0.779 | 0.712 | 0.079 | 0.610 | 0.781 |
130
  | Qwen3-VL-30B-A3B | 0.589 | 0.802 | 0.688 | 0.053 | 0.661 | 0.743 |
131
  | gpt-4.1 | 0.587 | 0.709 | 0.693 | 0.086 | 0.662 | 0.784 |
 
132
  | Qwen3-VL-30B-A3B-FP8 | 0.583 | 0.798 | 0.683 | 0.056 | 0.638 | 0.740 |
133
  | Qwen2.5-VL-32B | 0.577 | 0.767 | 0.649 | 0.232 | 0.493 | 0.743 |
134
  | gpt-5 (responses) | 0.573 | 0.746 | 0.650 | 0.080 | 0.687 | 0.704 |
135
  | Qwen2.5-VL-7B | 0.549 | 0.779 | 0.704 | 0.185 | 0.426 | 0.651 |
136
- | Qwen3-VL-4B-Instruct | 0.535 | 0.707 | 0.685 | 0.063 | 0.546 | 0.675 |
137
  | gpt-4.1-nano | 0.503 | 0.676 | 0.672 | 0.028 | 0.567 | 0.573 |
138
  | gpt-5-nano | 0.503 | 0.487 | 0.583 | 0.091 | 0.661 | 0.693 |
 
139
  | Qwen2.5-VL-3B | 0.402 | 0.613 | 0.654 | 0.045 | 0.203 | 0.494 |
140
- | Pixtral-12B-2409 | 0.342 | 0.327 | 0.555 | 0.026 | 0.325 | 0.475 |
141
 
 
 
 
 
 
142
 
143
  ## 💻 Usage Example
144
 
 
3
  dataset_name: mws-vision-bench
4
  language:
5
  - ru
6
+ license: cc-by-4.0
7
  tags:
8
  - benchmark
9
  - multimodal
 
114
 
115
  ---
116
 
117
+ ## 📊 Leaderboard (Validation Set)
118
+
119
+ Top models evaluated on this validation dataset:
120
 
121
  | Model | Overall | img→text | img→markdown | Grounding | KIE (JSON) | VQA |
122
  |-------|---------|----------|--------------|-----------|------------|-----|
 
124
  | **Gemini-2.5-flash** | **0.644** | 0.796 | 0.683 | 0.067 | 0.841 | 0.833 |
125
  | **gpt-4.1-mini** | **0.643** | 0.866 | 0.724 | 0.091 | 0.750 | 0.782 |
126
  | **Claude-4.5-Sonnet** | **0.639** | 0.723 | 0.676 | 0.377 | 0.728 | 0.692 |
127
+ | **gpt-5-mini** | **0.632** | 0.797 | 0.678 | 0.126 | 0.784 | 0.776 |
 
128
  | Qwen2.5-VL-72B | 0.631 | 0.848 | 0.712 | 0.220 | 0.644 | 0.732 |
129
  | gpt-5-mini (responses) | 0.594 | 0.743 | 0.567 | 0.118 | 0.811 | 0.731 |
 
130
  | Qwen3-VL-30B-A3B | 0.589 | 0.802 | 0.688 | 0.053 | 0.661 | 0.743 |
131
  | gpt-4.1 | 0.587 | 0.709 | 0.693 | 0.086 | 0.662 | 0.784 |
132
+ | Qwen3-VL-32B | 0.585 | 0.732 | 0.646 | 0.054 | 0.724 | 0.770 |
133
  | Qwen3-VL-30B-A3B-FP8 | 0.583 | 0.798 | 0.683 | 0.056 | 0.638 | 0.740 |
134
  | Qwen2.5-VL-32B | 0.577 | 0.767 | 0.649 | 0.232 | 0.493 | 0.743 |
135
  | gpt-5 (responses) | 0.573 | 0.746 | 0.650 | 0.080 | 0.687 | 0.704 |
136
  | Qwen2.5-VL-7B | 0.549 | 0.779 | 0.704 | 0.185 | 0.426 | 0.651 |
 
137
  | gpt-4.1-nano | 0.503 | 0.676 | 0.672 | 0.028 | 0.567 | 0.573 |
138
  | gpt-5-nano | 0.503 | 0.487 | 0.583 | 0.091 | 0.661 | 0.693 |
139
+ | Qwen3-VL-2B | 0.439 | 0.592 | 0.613 | 0.029 | 0.356 | 0.605 |
140
  | Qwen2.5-VL-3B | 0.402 | 0.613 | 0.654 | 0.045 | 0.203 | 0.494 |
 
141
 
142
+ *Scale: 0.0 - 1.0 (higher is better)*
143
+
144
+ **📝 Submit your model**: To evaluate on the private test set, contact [[email protected]](mailto:[email protected])
145
+
146
+ ---
147
 
148
  ## 💻 Usage Example
149