Multimodal Reasoning on Vision-Dependent Multimodal Reasoning Aggregate
59.82AVG ScorePAPO_D
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PAPO_DBackbone=Qwen2.5-VL, Model Size=7B2025.07 | 59.82 | 19.09 | |
| PAPO_GBackbone=Qwen2.5-VL, Model Size=7B2025.07 | 59.37 | 7.96 | |
| GRPOBackbone=Qwen2.5-VL, Model Size=7B2025.07 | 54.11 | — | |
| PAPO_DBackbone=Qwen2.5-VL, Model Size=3B2025.07 | 52.93 | 5.97 | |
| DAPOBackbone=Qwen2.5-VL, Model Size=7B2025.07 | 51.79 | — | |
| DAPOBackbone=Qwen2.5-VL, Model Size=3B2025.07 | 49.11 | — | |
| PAPO_GBackbone=Qwen3-VL (thinking), Model Size=2B2025.07 | 46.73 | 6.27 | |
| PAPO_GBackbone=Qwen2.5-VL, Model Size=3B2025.07 | 45.57 | 5.6 | |
| GRPOBackbone=Qwen3-VL (thinking), Model Size=2B2025.07 | 43.97 | — | |
| GRPOBackbone=Qwen2.5-VL, Model Size=3B2025.07 | 42.97 | — |