List your company

General Multimodal Reasoning on NaturalBench

78.62General ScoreQwen2.5-VL-7B + SRPO
72.01673.730575.44577.1595May 8, 2026

Evaluation results

MethodLinks
Qwen2.5-VL-7B + SRPOBackbone=Qwen2.5-VL-7B2026.0578.62
VL-Rethinker-7BBackbone=Qwen2.5-VL-7B2026.0578.17
PAPO-D-7BBackbone=Qwen2.5-VL-7B2026.0578.11
ThinkLite-VL-7BBackbone=Qwen2.5-VL-7B2026.0578.09
Qwen2.5-VL-7B + DAPOBackbone=Qwen2.5-VL-7B2026.0578.05
Vision-Matters-7BBackbone=Qwen2.5-VL-7B2026.0578.01
Perception-R1-7BBackbone=Qwen2.5-VL-7B2026.0577.96
VPPO-7BBackbone=Qwen2.5-VL-7B2026.0577.94
PAPO-G-7BBackbone=Qwen2.5-VL-7B2026.0577.92
Qwen2.5-VL-7B + GRPOBackbone=Qwen2.5-VL-7B2026.0577.92
PRCO-7BBackbone=Qwen2.5-VL-7B2026.0577.83
Qwen2.5-VL-7BBackbone=Qwen2.5-VL-7B2026.0577.67
Vision-SR1-7BBackbone=Qwen2.5-VL-7B2026.0577.65
Qwen2.5-VL-3B + SRPOBackbone=Qwen2.5-VL-3B2026.0574.83
PAPO-G-3BBackbone=Qwen2.5-VL-3B2026.0574.76
Qwen2.5-VL-3B + DAPOBackbone=Qwen2.5-VL-3B2026.0574.61
PAPO-D-3BBackbone=Qwen2.5-VL-3B2026.0574.45
PRCO-3BBackbone=Qwen2.5-VL-3B2026.0574.31
Qwen2.5-VL-3B + GRPOBackbone=Qwen2.5-VL-3B2026.0574.11
Vision-SR1-3BBackbone=Qwen2.5-VL-3B2026.0573.83
Qwen2.5-VL-3BBackbone=Qwen2.5-VL-3B2026.0572.27