Mathematical Reasoning on BRUMO 2025 (avg@32)
63.3Avg@32 ScoreGSPO
Evaluation Results
| Method | Links | |
|---|---|---|
| GSPOBase Model=Qwen3-14B, Reasoning Setting=Tool-integrated2026.05 | 63.3 | |
| CTPOBase Model=Qwen3-14B, Reasoning Setting=Tool-integrated2026.05 | 63 | |
| GRPOBase Model=Qwen3-14B, Reasoning Setting=Tool-integrated2026.05 | 61.9 | |
| CTPOBase Model=Qwen3-4B, Reasoning Setting=Tool-integrated2026.05 | 59.7 | |
| GSPOBase Model=Qwen3-4B, Reasoning Setting=Tool-integrated2026.05 | 56.8 | |
| GRPOBase Model=Qwen3-4B, Reasoning Setting=Tool-integrated2026.05 | 53.1 | |
| BaseBase Model=Qwen3-14B, Reasoning Setting=Tool-integrated2026.05 | 7 | |
| BaseBase Model=Qwen3-4B, Reasoning Setting=Tool-integrated2026.05 | 6.3 |