| Benchmark | Model | Rank | Results |
|---|---|---|---|
| mathematical-reasoning-on-aime-2025 | R²-OPD (Qwen3-1.7B) | – | Accuracy: 25.83 |
| mathematical-reasoning-on-aime24 | R²-OPD (DeepSeek-R1-Distill-Qwen-1.5B) | – | Acc: 32.5 |
| mathematical-reasoning-on-olympiadbench | R²-OPD (Qwen3-1.7B) | – | Accuracy: 54.31 |