Secure Coding on SecCodeBench
61.2Generation Success (w/o Hint)Qwen3-Coder-Next
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Qwen3-Coder-Nextevaluation_method=pass@k, weighting=security severity weighting2026.02 | 61.2 | 69.5 | 76.4 | 83.7 | |
| Claude-Opus-4-5evaluation_method=pass@k, weighting=security severity weighting2026.02 | 52.5 | 73.2 | 75.2 | 83.9 | |
| Claude-sonnet-4-5evaluation_method=pass@k, weighting=security severity weighting2026.02 | 43.4 | 57.6 | 62.2 | 76.4 | |
| Deepseek-V3.2evaluation_method=pass@k, weighting=security severity weighting2026.02 | 43.1 | 50.2 | 50 | 65.8 | |
| GLM-4.7evaluation_method=pass@k, weighting=security severity weighting2026.02 | 29.4 | 56 | 49.8 | 64.9 |