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

Multimodal Mathematical Reasoning on DynaMath

66.4DynaMath Score

Vision-R1-72B*

Updated 23d ago

Evaluation Results

Method	Links
Vision-R1-72B* 2025.03		66.4
Vision-R1-32B* 2025.03		65.6
GPT-4o 2025.03		64.9
Claude-3.5 Sonnet 2025.03		62.5
Qwen2.5-VL-72B 2025.03		61.2
PRCO-7B 2026.05		59.36
Qwen2.5-VL-7B + SRPO 2026.05		58.22
VPPO-7B 2026.05		56.88
PAPO-D-7B 2026.05		56.54
Vision-R1-7B 2025.03		56.3
Qwen2.5-VL-32B 2025.03		55.5
Qwen2.5-VL-7B + DAPO 2026.05		55.46
Perception-R1-7B 2026.05		55.36
VL-Rethinker-7B 2026.05		55.14
Vision-Matters-7B 2026.05		55.09
Qwen2.5-VL-7B + GRPO 2026.05		55.09
PAPO-G-7B 2026.05		53.11
Vision-SR1-7B 2026.05		52.71
ThinkLite-VL-7B 2026.05		51.79
Qwen2.5-VL-7B 2026.05		51.29
Qwen2.5-VL-7B 2025.03		50.7
Qwen2.5-VL-3B + SRPO 2026.05		48.54
PRCO-3B 2026.05		46.31
Vision-SR1-3B 2026.05		45.34
PAPO-D-3B 2026.05		45.16
Qwen2.5-VL-3B + GRPO 2026.05		45.06
PAPO-G-3B 2026.05		44.89
Qwen2.5-VL-3B + DAPO 2026.05		44.53
Qwen2.5-VL-3B 2026.05		43.15
InternVL2.5-78B 2025.03		41.4
LLaVA-CoT-11B 2025.03		34.6