| Benchmark | Model | Rank | Results |
|---|---|---|---|
| code-generation-on-humaneval | Think-Anywhere (Qwen2.5-Coder-7B-Instruct) | #25 | Pass@1: 91.5 |
| code-generation-on-mbpp | Think-Anywhere (Qwen2.5-Coder-7B-Instruct) | #25 | Accuracy: 82.9 |
| mathematical-reasoning-on-aime24 | Think-Anywhere | #124 | Acc: 17.3 |
| mathematical-reasoning-on-aime-2025 | Think-Anywhere | #131 | Accuracy: 17.7 |