Math Reasoning on AIME24 (mean@12)
77.5Mean@12 ScoreRLCSD
Evaluation results
| Method | Links | |
|---|---|---|
| RLCSDModel=Qwen3-8B2026.06 | 77.5 | |
| OPSDModel=Qwen3-8B2026.06 | 76.9 | |
| GRPOModel=Qwen3-8B2026.06 | 76.1 | |
| SDPOModel=Qwen3-8B2026.06 | 76.1 | |
| RLCSDModel=Olmo-3-7B2026.06 | 76.1 | |
| GRPOModel=Qwen3-4B2026.06 | 75.8 | |
| GRPOModel=Olmo-3-7B2026.06 | 75.8 | |
| OPSDModel=Olmo-3-7B2026.06 | 75.6 | |
| RLSDModel=Qwen3-8B2026.06 | 75.5 | |
| SRPOModel=Qwen3-8B2026.06 | 75.3 | |
| SRPOModel=Olmo-3-7B2026.06 | 75 | |
| RLSDModel=Olmo-3-7B2026.06 | 74.7 | |
| RLCSDModel=Qwen3-4B2026.06 | 74.4 | |
| OPSDModel=Qwen3-4B2026.06 | 74.2 | |
| BaseModel=Qwen3-8B2026.06 | 74.2 | |
| SDPOModel=Olmo-3-7B2026.06 | 74.2 | |
| BaseModel=Olmo-3-7B2026.06 | 73.9 | |
| BaseModel=Qwen3-4B2026.06 | 72.5 | |
| SRPOModel=Qwen3-4B2026.06 | 71.4 | |
| RLSDModel=Qwen3-4B2026.06 | 71.2 | |
| SDPOModel=Qwen3-4B2026.06 | 68.3 | |
| RLCSDModel=Qwen3-1.7B2026.06 | 53.1 | |
| GRPOModel=Qwen3-1.7B2026.06 | 51.6 | |
| OPSDModel=Qwen3-1.7B2026.06 | 50.8 | |
| BaseModel=Qwen3-1.7B2026.06 | 48.3 | |
| RLSDModel=Qwen3-1.7B2026.06 | 46.1 | |
| SRPOModel=Qwen3-1.7B2026.06 | 43.6 | |
| SDPOModel=Qwen3-1.7B2026.06 | 42.2 |