| Benchmark | Model | Rank | Results |
|---|---|---|---|
| mathematical-reasoning-on-aime24 | Qwen2.5-7B-Instruct with SA-MRPO (3 objectives) (extra data) | – | Acc: 16.5 |
| mathematical-reasoning-on-math-500 | Qwen2.5-7B-Instruct with SA-MRPO (3 objectives) (extra data) | – | Accuracy: 67.7 |
| mathematical-reasoning-on-minerva-math | Qwen2.5-7B-Instruct with SA-MRPO (3 objectives) (extra data) | – | Accuracy: 24.8 |
| mathematical-reasoning-on-olympiadbench | Qwen2.5-7B-Instruct with SA-MRPO (3 objectives) (extra data) | – | Accuracy: 26.1 |