| Benchmark | Model | Rank | Results |
|---|---|---|---|
| mathematical-reasoning-on-math-500 | TIP Soft-OR 50% (Qwen3-8B → Qwen3-4B) | #96 | Accuracy: 79.1 |
| mathematical-reasoning-on-aime24 | TIP Soft-OR 50% (Llama-3.3-70B → Llama-3.1-8B) | #108 | Acc: 26.0 |
| mathematical-reasoning-on-aime-2025 | TIP Soft-OR 20% (Qwen3-8B → Qwen3-4B) | #121 | Accuracy: 23.2 |