| Benchmark | Model | Rank | Results |
|---|---|---|---|
| code-generation-on-humaneval | TPPO (Qwen3.5-9B) | #8 | Pass@1: 96.5 |
| mathematical-reasoning-on-minerva-math | TPPO (Qwen3-4B) | #9 | Accuracy: 51.0 |
| mathematical-reasoning-on-olympiadbench | TPPO (Qwen3-4B) | #17 | Accuracy: 62.4 |
| code-generation-on-scicode | TPPO (Qwen3.5-9B) | #19 | Accuracy: 35.0 |
| mathematical-reasoning-on-math-500 | TPPO (Qwen3-4B) | #47 | Accuracy: 92.3 |
| code-generation-on-livecodebench-v6 | TPPO (Qwen3.5-9B) | #65 | Pass@1: 0.731 |