Mathematical Reasoning on AIME 25 (Acc@16, Tokens)
18.1Acc@16Generate
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GenerateModel=Qwen2.5-7B-Instruct2026.02 | 18.1 | 6,397 | |
| Self-VerifyModel=Qwen2.5-7B-Instruct2026.02 | 11.7 | 1,369 | |
| GenerateModel=Qwen2.5-3B-Instruct2026.02 | 8.1 | 4,744 | |
| Self-VerifyModel=Qwen2.5-3B-Instruct2026.02 | 6.3 | 2,252 | |
| Self-VerifyModel=Qwen2.5-1.5B-Instruct2026.02 | 1.3 | 1,545 | |
| GenerateModel=Qwen2.5-1.5B-Instruct2026.02 | 0.8 | 2,103 |