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

Mathematical Multimodal Reasoning on MM-Math

59.3Accuracy

Vision-R1-72B*

Updated 1mo ago

Evaluation Results

Method	Links
Vision-R1-72B* 2025.03		59.3
Vision-R1-32B* 2025.03		55.3
Qwen2.5-VL-72B 2025.03		45.6
Ours 2026.01		43.8
Ours [with DPS] 2026.01		43.4
Ours [with DPS and annealing] 2026.01		43.4
Vision-R1-7B 2025.03		40.2
VisonR1 7B 2026.01		40
VLAA-Thinker 7B 2026.01		39
Qwen2.5VL 7B 2026.01		36.4
MixedR1 7B 2026.01		35.8
Qwen2.5-VL-32B 2025.03		34.9
Qwen2.5-VL-7B 2025.03		34.1
R1-Onevision 7B 2026.01		32.9
GPT-4o 2026.01		31.8
GPT-4o 2025.03		31.8
Vision-R1-LlamaV-CI-11B 2025.03		26.1
Mulberry 2026.01		23.7
Mulberry-7B 2025.03		23.7
GPT-4V 2026.01		23.1
GPT-4V 2025.03		23.1
Mulberry-Llama-11B 2025.03		18.7
InternVL2.5-78B 2025.03		17.8
LLaVA-Cot-11B 2025.03		16.5
LLaVA-CoT-11B 2025.03		16.5
Llama-3.2-11B-V 2025.03		4.1