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

Visual Reasoning on REASONMAP-PLUS

88.95Weighted Accuracy

GPT-5

Updated 1mo ago

Evaluation Results

Method	Links
GPT-5 2025.10		88.95	86.4	91.6
Qwen2.5-VL-7B-Instruct + REWARDMAP 2025.10		74.25	72.18	76.42
Seed1.5-VL 2025.10		73.58	65.26	82.23
Qwen2.5-VL-7B-Instruct + RL (baseline) 2025.10		67.61	68.37	66.82
GPT-4o 2025.10		64.42	59.28	69.77
Qwen2.5-VL-7B-Instruct + SFT -> RL 2025.10		60.53	55.38	65.9
Qwen2.5-VL-32B-Instruct 2025.10		58.32	46.96	70.14
Qwen2.5-VL-7B-Instruct + SFT 2025.10		57.93	50.73	65.44
Qwen2.5-VL-72B-Instruct 2025.10		53.21	43.46	63.36
Qwen2.5-VL-7B-Instruct + RL (ReMax) 2025.10		45.39	38.37	52.7
Qwen2.5-VL-7B-Instruct + RL (GRPO) 2025.10		44.64	37.57	52.01
Qwen2.5-VL-7B-Instruct + RL (REINFORCE++) 2025.10		44.64	36.82	52.79
Qwen2.5-VL-7B-Instruct 2025.10		44.21	37.39	51.32
Qwen2.5-VL-3B-Instruct 2025.10		37.61	22.68	53.16
Kimi-VL-A3B-Thinking 2025.10		33.95	18.17	50.39
Kimi-VL-A3B-Instruct 2025.10		32.55	14.75	51.08