GRPO + CBS*
| Method | Links | |||
|---|---|---|---|---|
2026.02 | 46.63 | - | 5.2 | |
2026.02 | 45.63 | - | -1.7 | |
2026.02 | 43.92 | - | -0.4 | |
2026.02 | 42.41 | - | 2.9 | |
2026.01 | 42.17 | 4.56 | - | |
2026.01 | 41.04 | 3.43 | - | |
2026.01 | 39.79 | 2.18 | - | |
2026.01 | 38.26 | 0.65 | - | |
2026.01 | 37.94 | 0.33 | - | |
2026.01 | 37.93 | 0.32 | - | |
2026.01 | 37.71 | 0.1 | - | |
2026.01 | 37.61 | - | - | |
2026.01 | 37.4 | -0.21 | - | |
2026.02 | 32.41 | - | 5 | |
2026.02 | 31.08 | - | 7.5 | |
2026.02 | 30.17 | - | 4.4 | |
2026.02 | 30 | - | 15.7 | |
2026.01 | 22.04 | - | - |