Standard Mathematics on MATH 500 (Accuracy@4)
87.4Accuracy@4PAPO
Evaluation Results
| Method | Links | |
|---|---|---|
| PAPOModel=Qwen2.5-14B2026.03 | 87.4 | |
| PAPOModel=Qwen3-4B-Base2026.03 | 84.6 | |
| PAPOModel=Qwen2.5-7B2026.03 | 82.3 | |
| ORM(GRPO)Model=Qwen2.5-14B2026.03 | 82.3 | |
| ORM(DAPO)Model=Qwen3-4B-Base2026.03 | 82 | |
| ORM(GRPO)Model=Qwen2.5-7B2026.03 | 80.2 | |
| PAPOModel=Qwen2.5-3B2026.03 | 73.2 | |
| BaseModel=Qwen3-4B-Base2026.03 | 70.9 | |
| ORM(GRPO)Model=Qwen2.5-3B2026.03 | 69 | |
| BaseModel=Qwen2.5-14B2026.03 | 68.5 | |
| BaseModel=Qwen2.5-7B2026.03 | 67.3 | |
| BaseModel=Qwen2.5-3B2026.03 | 54.4 |