General Reasoning on GPQA Diamond (Pass@1, #Tok)
37.18Pass@1 AccuracySAS
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SASmax context length=4K, ratio r=0.32026.04 | 37.18 | 2,998 | |
| LAPO-Imax context length=4K2026.04 | 36.17 | 3,404 | |
| L1-Maxmax context length=4K2026.04 | 36.05 | 3,878 | |
| ThinkPrune-4kmax context length=4K2026.04 | 35.83 | 3,278 | |
| DeepScaleRmax context length=4K2026.04 | 35.7 | 4,860 | |
| GRPO-4Kmax context length=4K2026.04 | 32.48 | 1,515 |