Share your thoughts, 1 month free Claude Pro on usSee more

Visual Reasoning on REASONMAP Short questions

0.5998Weighted Accuracy

GPT-5

Updated 1mo ago

Evaluation Results

Method	Links
GPT-5 2025.10		0.5998	0.0948
GPT-4o 2025.10		0.4115	0.0684
Seed1.5-VL 2025.10		0.342	0.0525
Qwen2.5-VL-7B-Instruct + REWARDMAP 2025.10		0.3151	0.0621
Qwen2.5-VL-7B-Instruct + RL (baseline) 2025.10		0.2951	0.06
Qwen2.5-VL-7B-Instruct + SFT -> RL 2025.10		0.2882	0.0588
Qwen2.5-VL-7B-Instruct + RL (REINFORCE++) 2025.10		0.2717	0.0568
Qwen2.5-VL-72B-Instruct 2025.10		0.2665	0.0509
Qwen2.5-VL-7B-Instruct + RL (GRPO) 2025.10		0.2622	0.0552
Qwen2.5-VL-7B-Instruct + RL (ReMax) 2025.10		0.2622	0.0557
Qwen2.5-VL-32B-Instruct 2025.10		0.1649	0.0388
Qwen2.5-VL-7B-Instruct + SFT 2025.10		0.1363	0.0409
Qwen2.5-VL-7B-Instruct 2025.10		0.1328	0.0401
Kimi-VL-A3B-Instruct 2025.10		0.1276	0.033
Qwen2.5-VL-3B-Instruct 2025.10		0.0868	0.0275
Kimi-VL-A3B-Thinking 2025.10		0.0547	0.0244