Code Reasoning on OJBench
10.34AccuracyDFT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DFTModel=Qwen3 32B2025.10 | 10.34 | 10.34 | |
| SFTModel=Qwen3 32B2025.10 | 9.91 | 9.91 | |
| RandomModel=Qwen3 32B2025.10 | 9.91 | 9.91 | |
| VCOREModel=Qwen3 32B2025.10 | 9.48 | 9.48 | |
| iw-SFTModel=Qwen3 32B2025.10 | 6.9 | 6.9 | |
| DFTModel=Qwen3 8B2025.10 | 6.03 | 6.03 | |
| DFTModel=Qwen3 4B2025.10 | 5.6 | 5.6 | |
| iw-SFTModel=Qwen3 8B2025.10 | 5.6 | 5.6 | |
| OriginalModel=Qwen3 32B2025.10 | 5.6 | 5.6 | |
| SFTModel=Qwen3 8B2025.10 | 4.74 | 4.74 | |
| VCOREModel=Qwen3 8B2025.10 | 4.74 | 4.74 | |
| OriginalModel=Qwen3 8B2025.10 | 4.31 | 4.31 | |
| RandomModel=Qwen3 8B2025.10 | 4.31 | 4.31 | |
| VCOREModel=Qwen3 4B2025.10 | 3.88 | 3.88 | |
| RandomModel=LLaMA3.1 8B Instruct2025.10 | 3.02 | 3.02 | |
| OriginalModel=Qwen3 4B2025.10 | 3.02 | 3.02 | |
| iw-SFTModel=Qwen3 4B2025.10 | 2.59 | 2.59 | |
| VCOREModel=LLaMA3.1 8B Instruct2025.10 | 2.58 | 2.58 | |
| SFTModel=Qwen3 4B2025.10 | 2.16 | 2.16 | |
| OriginalModel=LLaMA3.1 8B Instruct2025.10 | 1.29 | 1.29 | |
| SFTModel=LLaMA3.1 8B Instruct2025.10 | 0.86 | 0.86 | |
| iw-SFTModel=LLaMA3.1 8B Instruct2025.10 | 0.86 | 0.86 | |
| RandomModel=Qwen3 4B2025.10 | 0.86 | 0.86 | |
| DFTModel=LLaMA3.1 8B Instruct2025.10 | 0 | 0 |