| Benchmark | Model | Rank | Results |
|---|---|---|---|
| mathematical-reasoning-on-aime-2025 | Qwen3-4B-Non-Thinking + IDA-OPD | – | Accuracy: 53.3 |
| mathematical-reasoning-on-aime24 | Qwen3-8B-Non-Thinking + IDA-OPD | – | Acc: 63.3 |
| mathematical-reasoning-on-hmmt-feb-2025 | Qwen3-8B-Non-Thinking + IDA-OPD | – | Accuracy: 37.6 |