Mathematical Reasoning on MATH500 (mean@8)
84.6Mean@8 ScoreACPO
Evaluation Results
| Method | Links | |
|---|---|---|
| ACPOModel=DeepSeek-R1-Distill-Qwen-7B2026.07 | 84.6 | |
| ACPOModel=Qwen3-8B-Base2026.07 | 83.6 | |
| SAPOModel=DeepSeek-R1-Distill-Qwen-7B2026.07 | 83.2 | |
| SAPOModel=Qwen3-8B-Base2026.07 | 82.4 | |
| GTPOModel=Qwen3-8B-Base2026.07 | 81.62 | |
| DAPOModel=DeepSeek-R1-Distill-Qwen-7B2026.07 | 80.4 | |
| GTPOModel=DeepSeek-R1-Distill-Qwen-7B2026.07 | 78.85 | |
| 80/20Model=DeepSeek-R1-Distill-Qwen-7B2026.07 | 78.72 | |
| 80/20Model=Qwen3-8B-Base2026.07 | 78.22 | |
| GRPOModel=DeepSeek-R1-Distill-Qwen-7B2026.07 | 77.95 | |
| ACPOModel=Qwen2.5-Math-7B2026.07 | 77.8 | |
| DAPOModel=Qwen3-8B-Base2026.07 | 76.8 | |
| SAPOModel=Qwen2.5-Math-7B2026.07 | 76.4 | |
| GTPOModel=Qwen2.5-Math-7B2026.07 | 75.5 | |
| DAPOModel=Qwen2.5-Math-7B2026.07 | 75 | |
| 80/20Model=Qwen2.5-Math-7B2026.07 | 74.75 | |
| GRPOModel=Qwen3-8B-Base2026.07 | 72.8 | |
| GRPOModel=Qwen2.5-Math-7B2026.07 | 71.87 |