List your company
ResearchFormerly WizWand

Visual Reasoning on A-OKVQA

5.4ECEAvg(VF, Contr.)
4.28811.79419.326.806Sep 30, 2025
Updated 6mo ago

Evaluation results

MethodLinks
Avg(VF, Contr.)Model=Qwen2.52025.095.4
Avg(VF, Contr.)Model=Avg2025.098.4
Avg(VF, Contr.)Model=GPT-4o2025.099
Visual FidelityModel=GPT-4o2025.099.9
Avg(VF, Contr.)Model=LLaVA2025.0910.9
ContrastivenessModel=Qwen2.52025.0913.6
Visual FidelityModel=Qwen2.52025.0913.7
Visual FidelityModel=Avg2025.0914.4
Min(VF, Contr.)Model=Qwen2.52025.0914.4
Min(VF, Contr.)Model=LLaVA2025.0914.7
InformativenessModel=Qwen2.52025.0915.2
Prod(VF, Contr.)Model=Qwen2.52025.0915.4
PlausibilityModel=LLaVA2025.0916.2
Prod(VF, Contr.)Model=LLaVA2025.0916.4
Min(VF, Contr.)Model=Avg2025.0917.3
ContrastivenessModel=Avg2025.0917.4
ContrastivenessModel=LLaVA2025.0917.6
Prod(VF, Contr.)Model=Avg2025.0918.4
Visual FidelityModel=LLaVA2025.0919.7
ContrastivenessModel=GPT-4o2025.0921
SimulatabilityModel=Qwen2.52025.0921.9
Min(VF, Contr.)Model=GPT-4o2025.0922.7
InformativenessModel=GPT-4o2025.0922.8
Prod(VF, Contr.)Model=GPT-4o2025.0923.3
InformativenessModel=Avg2025.0923.7
PlausibilityModel=Avg2025.0925
SimulatabilityModel=Avg2025.0927
PlausibilityModel=Qwen2.52025.0927
SimulatabilityModel=LLaVA2025.0928.8
SimulatabilityModel=GPT-4o2025.0930.2
PlausibilityModel=GPT-4o2025.0931.7
InformativenessModel=LLaVA2025.0933.2