Code Verification on HumanEval+
87.05Pass@1CVeDRL
Evaluation results
| Method | Links | |
|---|---|---|
| CVeDRLScale=0.6B, Policy Model=GPT-4o-m2026.01 | 87.05 | |
| Qwen3Scale=32B, Policy Model=GPT-4o-m2026.01 | 86.58 | |
| CodeRMScale=8.0B, Policy Model=GPT-4o-m2026.01 | 86.49 | |
| CodeTScale=-, Policy Model=GPT-4o-m2026.01 | 85.52 | |
| Llama3.1Scale=70B, Policy Model=GPT-4o-m2026.01 | 85.48 | |
| MBR-EScale=-, Policy Model=GPT-4o-m2026.01 | 85.31 | |
| Qwen3Scale=0.6B, Policy Model=GPT-4o-m2026.01 | 83.01 | |
| VanillaScale=-, Policy Model=GPT-4o-m2026.01 | 82.57 | |
| Llama3.1Scale=70B, Policy Model=GPT-3.52026.01 | 79.88 | |
| CVeDRLScale=0.6B, Policy Model=GPT-3.52026.01 | 78.96 | |
| CVeDRLScale=0.6B, Policy Model=Llama3 70B2026.01 | 78.72 | |
| CodeRMScale=8.0B, Policy Model=Llama3 70B2026.01 | 78.66 | |
| Qwen3Scale=32B, Policy Model=Llama3 70B2026.01 | 78.52 | |
| Qwen3Scale=32B, Policy Model=GPT-3.52026.01 | 78.45 | |
| Llama3.1Scale=70B, Policy Model=Llama3 70B2026.01 | 78.41 | |
| CodeRMScale=8.0B, Policy Model=GPT-3.52026.01 | 78.13 | |
| CodeTScale=-, Policy Model=Llama3 70B2026.01 | 76.14 | |
| MBR-EScale=-, Policy Model=Llama3 70B2026.01 | 75.47 | |
| CodeTScale=-, Policy Model=GPT-3.52026.01 | 73.92 | |
| Qwen3Scale=0.6B, Policy Model=Llama3 70B2026.01 | 73.91 | |
| VanillaScale=-, Policy Model=Llama3 70B2026.01 | 73.1 | |
| CVeDRLScale=0.6B, Policy Model=Llama3 8B2026.01 | 72.14 | |
| CodeRMScale=8.0B, Policy Model=Llama3 8B2026.01 | 72.13 | |
| Llama3.1Scale=70B, Policy Model=Llama3 8B2026.01 | 71.95 | |
| Qwen3Scale=32B, Policy Model=Llama3 8B2026.01 | 70.64 | |
| MBR-EScale=-, Policy Model=GPT-3.52026.01 | 70.53 | |
| Qwen3Scale=0.6B, Policy Model=GPT-3.52026.01 | 69.16 | |
| VanillaScale=-, Policy Model=GPT-3.52026.01 | 67.44 | |
| CodeTScale=-, Policy Model=Llama3 8B2026.01 | 65.22 | |
| MBR-EScale=-, Policy Model=Llama3 8B2026.01 | 60.18 | |
| Qwen3Scale=0.6B, Policy Model=Llama3 8B2026.01 | 55.75 | |
| VanillaScale=-, Policy Model=Llama3 8B2026.01 | 53.43 |