Competition Mathematics on AIME 2025 (Accuracy avg@4)
30.7Accuracy (avg@4)PAPO
Evaluation Results
| Method | Links | |
|---|---|---|
| PAPOModel=Qwen3-4B-Base2026.03 | 30.7 | |
| ORM(DAPO)Model=Qwen3-4B-Base2026.03 | 26 | |
| PAPOModel=Qwen2.5-14B2026.03 | 17.8 | |
| ORM(GRPO)Model=Qwen2.5-14B2026.03 | 13.8 | |
| PAPOModel=Qwen2.5-7B2026.03 | 13.1 | |
| ORM(GRPO)Model=Qwen2.5-7B2026.03 | 10.8 | |
| BaseModel=Qwen2.5-14B2026.03 | 8.3 | |
| BaseModel=Qwen2.5-7B2026.03 | 7.5 | |
| BaseModel=Qwen3-4B-Base2026.03 | 5 | |
| ORM(GRPO)Model=Qwen2.5-3B2026.03 | 4.2 | |
| PAPOModel=Qwen2.5-3B2026.03 | 3.3 | |
| BaseModel=Qwen2.5-3B2026.03 | 0.8 |