List your company
ResearchFormerly WizWand

General Multimodal Reasoning on MMMU-Pro

38.09ScoreQwen2.5-VL-7B + SRPO
25.578828.826932.07535.3231May 8, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Qwen2.5-VL-7B + SRPOBackbone=Qwen2.5-VL-7B2026.0538.09
PRCO-7BBackbone=Qwen2.5-VL-7B2026.0536.94
VPPO-7BBackbone=Qwen2.5-VL-7B2026.0535.89
PAPO-G-7BBackbone=Qwen2.5-VL-7B2026.0535.37
PAPO-D-7BBackbone=Qwen2.5-VL-7B2026.0535.2
Qwen2.5-VL-7B + DAPOBackbone=Qwen2.5-VL-7B2026.0535.2
Vision-Matters-7BBackbone=Qwen2.5-VL-7B2026.0534.91
Perception-R1-7BBackbone=Qwen2.5-VL-7B2026.0534.73
Vision-SR1-7BBackbone=Qwen2.5-VL-7B2026.0534.33
VL-Rethinker-7BBackbone=Qwen2.5-VL-7B2026.0533.35
Qwen2.5-VL-7B + GRPOBackbone=Qwen2.5-VL-7B2026.0532.65
ThinkLite-VL-7BBackbone=Qwen2.5-VL-7B2026.0531.33
Qwen2.5-VL-7BBackbone=Qwen2.5-VL-7B2026.0530.73
Qwen2.5-VL-3B + SRPOBackbone=Qwen2.5-VL-3B2026.0529.65
Vision-SR1-3BBackbone=Qwen2.5-VL-3B2026.0528.61
Qwen2.5-VL-3B + DAPOBackbone=Qwen2.5-VL-3B2026.0528.15
PAPO-D-3BBackbone=Qwen2.5-VL-3B2026.0527.91
PRCO-3BBackbone=Qwen2.5-VL-3B2026.0527.8
PAPO-G-3BBackbone=Qwen2.5-VL-3B2026.0527.68
Qwen2.5-VL-3B + GRPOBackbone=Qwen2.5-VL-3B2026.0527.45
Qwen2.5-VL-3BBackbone=Qwen2.5-VL-3B2026.0526.06