General Reasoning on CMMLU (test)
64.1AccuracyQwen2.5-7B-Inst.
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-7B-Inst.Parameters=7B2026.03 | 64.1 | |
| VEPO-7B (Full)Parameters=7B, Configuration=Full2026.03 | 63.2 | |
| VEPO-7B w/o CPTParameters=7B, CPT=excluded2026.03 | 62 | |
| VEPO-7B-SFTParameters=7B, Stage=SFT2026.03 | 61.4 | |
| Qwen3-8BParameters=8B2026.03 | 60.8 | |
| Tower-Plus-9BParameters=9B2026.03 | 57.2 | |
| Gemma3-27B-ITParameters=27B2026.03 | 55.9 | |
| Apertus-8B-Inst.Parameters=8B2026.03 | 45.3 |