CWE-bench v1: Security Misconfiguration
Code · 2026-09-25
Programmatic pass@1 on the OWASP 2025 A02 subset of CWE-bench v1; exact category denominator is not published.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5.5 | 75.0 |
| GPT-6 Astra | 75.0 |
| Grok 4.7 | 75.0 |
| Gemini 4 Argon | 66.0 |
| Claude Fable 5.1 | 64.0 |
| Hy4 Preview | 61.0 |
| DeepSeek V4.1 Flash | 57.0 |
| Grok 4.6 | 57.0 |
| Inkling | 55.0 |
| Muse Spark 1.3 | 55.0 |
| GPT-6 Sol | 43.0 |