Math Reasoning on Olympiad (Accuracy, Pass@1)
68.55AccuracyVCORE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| VCOREModel=Qwen3 32B2025.10 | 68.55 | 68.55 | |
| DFTModel=Qwen3 32B2025.10 | 67.66 | 67.66 | |
| VCOREModel=Qwen3 4B2025.10 | 66.17 | 66.17 | |
| DFTModel=Qwen3 8B2025.10 | 65.13 | 65.13 | |
| OriginalModel=Qwen3 32B2025.10 | 64.99 | 64.99 | |
| VCOREModel=Qwen3 8B2025.10 | 64.84 | 64.84 | |
| SFTModel=Qwen3 8B2025.10 | 63.8 | 63.8 | |
| SFTModel=Qwen3 32B2025.10 | 63.8 | 63.8 | |
| iw-SFTModel=Qwen3 32B2025.10 | 63.65 | 63.65 | |
| DFTModel=Qwen3 4B2025.10 | 62.91 | 62.91 | |
| iw-SFTModel=Qwen3 4B2025.10 | 62.31 | 62.31 | |
| RandomModel=Qwen3 32B2025.10 | 62.31 | 62.31 | |
| SFTModel=Qwen3 4B2025.10 | 61.72 | 61.72 | |
| iw-SFTModel=Qwen3 8B2025.10 | 61.72 | 61.72 | |
| RandomModel=Qwen3 8B2025.10 | 61.57 | 61.57 | |
| RandomModel=Qwen3 4B2025.10 | 60.98 | 60.98 | |
| OriginalModel=Qwen3 4B2025.10 | 60.09 | 60.09 | |
| OriginalModel=Qwen3 8B2025.10 | 60.09 | 60.09 | |
| VCOREModel=LLaMA3.1 8B Instruct2025.10 | 25.96 | 25.96 | |
| SFTModel=LLaMA3.1 8B Instruct2025.10 | 23.59 | 23.59 | |
| iw-SFTModel=LLaMA3.1 8B Instruct2025.10 | 22.55 | 22.55 | |
| RandomModel=LLaMA3.1 8B Instruct2025.10 | 22.55 | 22.55 | |
| DFTModel=LLaMA3.1 8B Instruct2025.10 | 18.84 | 18.84 | |
| OriginalModel=LLaMA3.1 8B Instruct2025.10 | 17.8 | 17.8 |