Mathematical Reasoning on AIME 2024 (pass@16, avg@16)
86.7Pass@16Self-ReSET
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Self-ReSETModel Family=DS-Qwen-7B2026.05 | 86.7 | 52.9 | |
| BaseModel Family=DS-Qwen-7B2026.05 | 83.3 | 51.9 | |
| RECAPModel Family=DS-Qwen-7B2026.05 | 83.3 | 52.1 | |
| DAPOModel Family=DS-Qwen-7B2026.05 | 83.3 | 50.2 | |
| Self-ReSETModel Family=Qwen3-8B2026.05 | 83.3 | 62.3 | |
| DAPOModel Family=DS-Llama-8B2026.05 | 80 | 46.5 | |
| BaseModel Family=Qwen3-8B2026.05 | 80 | 60 | |
| BaseModel Family=DS-Llama-8B2026.05 | 76.7 | 44.4 | |
| Self-ReSETModel Family=DS-Llama-8B2026.05 | 76.7 | 47.1 | |
| RECAPModel Family=Qwen3-8B2026.05 | 76.7 | 62.1 | |
| DAPOModel Family=Qwen3-8B2026.05 | 76.7 | 60 | |
| RECAPModel Family=DS-Llama-8B2026.05 | 73.3 | 43.5 |