List your company
ResearchFormerly WizWand

Outcome Reasoning on COCO

77.8M' (F1 Mean)GPT-5
48.1655.85563.5571.245May 17, 2025
Updated 8mo ago

Evaluation results

MethodLinks
GPT-5Model=GPT-52025.0577.870.1
GPT-o4Model=GPT-o42025.0575.366.9
Llama4-MModel=Llama4-M2025.0565.458.7
DeepSeekModel=DeepSeek2025.0561.354.6
Qwen3Model=Qwen32025.0560.153.4
Gemini2.5Model=Gemini2.52025.0557.851.5
Llama4-SModel=Llama4-S2025.0549.342.7