Mathematical Reasoning on Minerva (Avg, Pass metrics)
49.5Average ScoreDistIL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DistILModel=Qwen3-8B2026.06 | 49.5 | 58.4 | |
| SDPOModel=Qwen3-8B2026.06 | 49.2 | 57.7 | |
| OPSDModel=Qwen3-8B2026.06 | 49.2 | 57.8 | |
| BaseModel=Qwen3-8B2026.06 | 48.9 | 56.3 | |
| GRPOModel=Qwen3-8B2026.06 | 48.9 | 56.3 | |
| SFTModel=Qwen3-8B2026.06 | 48.2 | 56.2 | |
| DistILModel=Qwen3-4B2026.06 | 44.2 | 52.9 | |
| SDPOModel=Qwen3-4B2026.06 | 43.4 | 52 | |
| OPSDModel=Qwen3-4B2026.06 | 43.4 | 51.8 | |
| BaseModel=Qwen3-4B2026.06 | 43.2 | 48.6 | |
| GRPOModel=Qwen3-4B2026.06 | 43.2 | 48.6 | |
| SFTModel=Qwen3-4B2026.06 | 41.9 | 50.4 |