Math Reasoning on AIME 2025 (Pass@1, AES)
27.92Pass@1LAPO-I
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LAPO-I#Tok=52902026.04 | 27.92 | — | |
| DeepScaleR#Tok=64442026.04 | 26.88 | — | |
| ThinkPrune-4k#Tok=51772026.04 | 26.67 | — | |
| SAS#Tok=42952026.04 | 26.67 | — | |
| GRPO-4K#Tok=48122026.04 | 25.42 | — | |
| L1-Max#Tok=20322026.04 | 21.67 | — |