Multimodal Math Reasoning on MathVista (Score)
76.3ScoreQwen2.5-VL-7B + SRPO
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-VL-7B + SRPOBackbone=Qwen2.5-VL-7B2026.05 | 76.3 | |
| Qwen2.5-VL-7B + DAPOBackbone=Qwen2.5-VL-7B2026.05 | 74.5 | |
| PRCO-7BBackbone=Qwen2.5-VL-7B2026.05 | 74.1 | |
| VPPO-7BBackbone=Qwen2.5-VL-7B2026.05 | 73.6 | |
| PAPO-D-7BBackbone=Qwen2.5-VL-7B2026.05 | 73.3 | |
| ThinkLite-VL-7BBackbone=Qwen2.5-VL-7B2026.05 | 73.1 | |
| Perception-R1-7BBackbone=Qwen2.5-VL-7B2026.05 | 73.1 | |
| Vision-Matters-7BBackbone=Qwen2.5-VL-7B2026.05 | 73.1 | |
| VL-Rethinker-7BBackbone=Qwen2.5-VL-7B2026.05 | 72.6 | |
| Qwen2.5-VL-7B + GRPOBackbone=Qwen2.5-VL-7B2026.05 | 72.3 | |
| Vision-SR1-7BBackbone=Qwen2.5-VL-7B2026.05 | 71.8 | |
| PAPO-G-7BBackbone=Qwen2.5-VL-7B2026.05 | 71.5 | |
| Qwen2.5-VL-7BBackbone=Qwen2.5-VL-7B2026.05 | 69.8 | |
| PRCO-3BBackbone=Qwen2.5-VL-3B2026.05 | 67.9 | |
| Qwen2.5-VL-3B + SRPOBackbone=Qwen2.5-VL-3B2026.05 | 67.8 | |
| Qwen2.5-VL-3B + DAPOBackbone=Qwen2.5-VL-3B2026.05 | 67.6 | |
| PAPO-D-3BBackbone=Qwen2.5-VL-3B2026.05 | 65.8 | |
| Qwen2.5-VL-3B + GRPOBackbone=Qwen2.5-VL-3B2026.05 | 64.9 | |
| PAPO-G-3BBackbone=Qwen2.5-VL-3B2026.05 | 64.7 | |
| Qwen2.5-VL-3BBackbone=Qwen2.5-VL-3B2026.05 | 63.3 | |
| Vision-SR1-3BBackbone=Qwen2.5-VL-3B2026.05 | 62.2 |