| Benchmark | Model | Rank | Results |
|---|---|---|---|
| mathematical-reasoning-on-aime-2025 | A3PO (Deepseek-R1-Distill-Qwen-7B) | – | Accuracy: 54.1 |
| mathematical-reasoning-on-aime24 | A3PO (Deepseek-R1-Distill-Qwen-7B) | – | Acc: 65.2 |
| mathematical-reasoning-on-math-500 | A3PO (Deepseek-R1-Distill-Qwen-7B) | – | Accuracy: 96.9 |