Model,Ordinal (Win rate),Cardinal (Score),RO Stability,Stress,CFI,SRMR,RMSEA reka-flash-3,0.3676470588235294,0.4899723723251134,0.2885614355228928,0.21908343015733012,0.6356444444444445,0.32448055555555566,0.3265472222222222 Llama-3.3-70B-Instruct,0.7720588235294118,0.7826258982618856,0.8101440811870466,0.15086913865310292,0.7070694444444444,0.2484805555555556,0.2872888888888889 Llama-3.1-70B-Instruct,0.6732026143790849,0.7716257197180342,0.7627363693137972,0.1590258306220288,0.7405305555555556,0.2069805555555555,0.2524638888888888 Llama-3.1-Nemotron-70B-Instruct,0.7663398692810457,0.806561245024004,0.7986269216633632,0.15639272646142155,0.7771694444444445,0.1574805555555555,0.2053138888888889 Llama-3.1-8B-Instruct,0.4566993464052287,0.6200396929619997,0.5544135256446929,0.177589336660551,0.6203138888888889,0.3474166666666667,0.3526444444444444 Llama-3.2-3B-Instruct,0.35947712418300654,0.4920203960749596,0.3095420768898777,0.23477458340427637,0.6402583333333334,0.3671361111111112,0.34035277777777784 Llama-3.2-1B-Instruct,0.14950980392156865,0.28494630130657267,0.017845599375419068,0.29245365371576204,0.4570666666666667,0.49829999999999997,0.458125 Mistral-Large-Instruct-2411,0.5857843137254901,0.7334649247049873,0.7061158170869489,0.17181887070789792,0.7237666666666667,0.24669166666666675,0.262 Mistral-Large-Instruct-2407,0.6846405228758169,0.7865403071348558,0.7786865940656633,0.18242669696158476,0.7757583333333334,0.19862499999999994,0.21713055555555572 Mistral-Nemo-Instruct-2407,0.34068627450980393,0.5243749748919985,0.409719354830282,0.21289984240736382,0.565475,0.40046944444444443,0.39598333333333324 Mistral-Small-3.1-24B-Instruct-2503,0.5326797385620915,0.7026337346865648,0.6848141294613206,0.17721164035276304,0.6713749999999999,0.29685555555555565,0.3154944444444444 QwQ-32B,0.7230392156862745,0.7719313363889678,0.8091295835194909,0.1769481985217758,0.6934333333333333,0.27074999999999994,0.30680277777777776 Qwen2.5-VL-72B-Instruct,0.8619281045751634,0.8360251140700989,0.8125460511443046,0.15932218090531514,0.8454611111111111,0.10019722222222227,0.14792499999999997 Qwen2.5-VL-7B-Instruct,0.2450980392156863,0.42422904460473104,0.28582477835186304,0.25948620101404807,0.45279444444444444,0.4787000000000001,0.464075 Qwen2.5-VL-3B-Instruct,0.1111111111111111,0.2655381207197522,0.059657867275330144,0.2877042811210807,0.3650055555555556,0.6013444444444445,0.5902833333333335 Qwen2.5-14B-Instruct-1M,0.5359477124183007,0.7022023055217502,0.6551137143166985,0.1785253019816836,0.7130944444444444,0.26870277777777785,0.26870277777777773 Dracarys2-72B-Instruct,0.7295751633986929,0.7740958932029343,0.789501612210195,0.15836985877285903,0.7307833333333333,0.2581972222222222,0.27945555555555546 dummy,0.09967320261437908,0.2291015386716794,-0.009004148398032956,0.2928877637010999,0.3755222222222222,0.622275,0.5915305555555557