Mathematical Reasoning on AIME 25, AIME 26, BRUMO 25, HMMT 25 Aggregate
58.8Average Score @ 32CTPO
Evaluation Results
| Method | Links | |
|---|---|---|
| CTPOBase Model=Qwen3-14B, Reasoning Setting=Tool-integrated2026.05 | 58.8 | |
| GSPOBase Model=Qwen3-14B, Reasoning Setting=Tool-integrated2026.05 | 55.5 | |
| GRPOBase Model=Qwen3-14B, Reasoning Setting=Tool-integrated2026.05 | 54.6 | |
| CTPOBase Model=Qwen3-4B, Reasoning Setting=Tool-integrated2026.05 | 51.4 | |
| GSPOBase Model=Qwen3-4B, Reasoning Setting=Tool-integrated2026.05 | 47.7 | |
| GRPOBase Model=Qwen3-4B, Reasoning Setting=Tool-integrated2026.05 | 43.2 | |
| BaseBase Model=Qwen3-14B, Reasoning Setting=Tool-integrated2026.05 | 6.1 | |
| BaseBase Model=Qwen3-4B, Reasoning Setting=Tool-integrated2026.05 | 3.9 |