Logic on KORBench
56.2AccuracyQwen2.5-32B
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen2.5-32BParameters=32B2026.05 | 56.2 | |
| Qwen2.5-72BParameters=72B2026.05 | 53 | |
| Qwen2.5-14BParameters=14B2026.05 | 50.6 | |
| FORGE-7B-NP-MATHParameters=7B, Configuration=NP-MATH2026.05 | 46.6 | |
| FORGE-7B-MATHParameters=7B, Configuration=MATH2026.05 | 45.9 | |
| LLama3.1-8bParameters=8B2026.05 | 44.5 | |
| FORGE-7B-NPParameters=7B, Configuration=NP2026.05 | 44.1 | |
| Qwen2.5-7BParameters=7B2026.05 | 42.9 | |
| InternLM3-8BParameters=8B2026.05 | 40.7 | |
| Qwen2.5-3BParameters=3B2026.05 | 36.6 | |
| Qwen2.5-7B-SFTParameters=7B, Configuration=SFT2026.05 | 30.1 |