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

General Multimodal Reasoning on MMMU (val)

71.33Pass@1 Accuracy

Qwen3-VL-8B-DeepVision

Updated 1mo ago

Evaluation Results

Method	Links
Qwen3-VL-8B-DeepVision 2026.02		71.33
MiMo-VL-7B-DeepVision 2026.02		71
GPT-5-Nano-High 2026.02		70.78
Qwen3-VL-8B-Thinking 2026.02		69.33
MiMo-VL-7B-MM-Eureka 2026.02		67.67
Qwen3-VL-8B-Instruct 2026.02		67.66
MiMo-VL-7B-RL-2508 2026.02		67.44
MiMo-VL-7B-OpenMMReasoner 2026.02		66.78
MiMo-VL-7B-MathBook 2026.02		66.33
Gemini-2.5-Flash-Lite 2026.02		64.77
MiMo-VL-7B-SFT-2508 2026.02		63.77