Our new X account is live! Follow @wizwand_team for updates

Visual Reasoning on REASONMAP Long questions

62.5Weighted Accuracy

GPT-5

Updated 3d ago

Evaluation Results

Method	Links
GPT-5 2025.10		62.5	19.75
GPT-4o 2025.10		42.8	13.57
Seed1.5-VL 2025.10		38.02	11.96
Qwen2.5-VL-7B-Instruct + REWARDMAP 2025.10		31.77	11.22
Qwen2.5-VL-7B-Instruct + SFT -> RL 2025.10		30.38	10.62
Qwen2.5-VL-7B-Instruct + RL (baseline) 2025.10		29.51	10.41
Qwen2.5-VL-7B-Instruct + RL (REINFORCE++) 2025.10		27.6	10.12
Qwen2.5-VL-7B-Instruct + RL (ReMax) 2025.10		27.26	9.99
Qwen2.5-VL-7B-Instruct + RL (GRPO) 2025.10		26.04	9.52
Qwen2.5-VL-72B-Instruct 2025.10		24.22	8.8
Qwen2.5-VL-32B-Instruct 2025.10		15.71	6.84
Kimi-VL-A3B-Instruct 2025.10		12.33	5.37
Qwen2.5-VL-7B-Instruct + SFT 2025.10		9.11	6.25
Qwen2.5-VL-3B-Instruct 2025.10		7.99	3.7
Qwen2.5-VL-7B-Instruct 2025.10		7.12	5.74
Kimi-VL-A3B-Thinking 2025.10		5.47	3.17