[Long-Horizon] Warm-Start SFT + RL (Outcome + Process)
| Method | Links | ||
|---|---|---|---|
| 65.2 | 79.9 | ||
2025.10 | 64.8 | 73.1 | |
2025.10 | 64.8 | 75.2 | |
| 64.5 | 79.1 | ||
| 63.5 | 77.3 | ||
| 63 | 75.2 | ||
| 62.8 | 74.7 | ||
| 62.7 | 76.6 | ||
| 62.1 | 73 | ||
| 62.1 | 73.9 | ||
| 60.5 | 72.2 | ||
2025.10 | 59.3 | 68.1 | |
2025.10 | 59 | 72.8 | |
2025.10 | 58.8 | 67 | |
| 57.6 | 69.9 | ||
2025.10 | 56.4 | - | |
2025.10 | 54.4 | 66.8 | |
2025.10 | 54.3 | 66.5 | |
2025.10 | 52.4 | 69.6 | |
2025.10 | 50.7 | 64 | |
2025.10 | 49.3 | 59.9 | |
2025.10 | 46 | 63.7 | |
| 45.6 | 60.3 | ||
2025.10 | 42 | 38.8 | |
2025.10 | 36 | 58.5 | |
| 32.5 | 69.9 | ||
2025.10 | 32.1 | 60.9 | |
2025.10 | 29.1 | 59.8 | |
| 16.3 | 22.6 |