CWE-bench v1: JS/TS
Code · 2026-09-25
Programmatic pass@1 across the 26 JavaScript/TypeScript tasks in CWE-bench v1, averaging four rollouts per task.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5.5 | 74.0 |
| GPT-6 Astra | 72.0 |
| Grok 4.7 | 66.0 |
| Gemini 4 Argon | 65.0 |
| Muse Spark 1.3 | 65.0 |
| Grok 4.6 | 63.0 |
| Claude Fable 5.1 | 59.0 |
| Hy4 Preview | 57.0 |
| DeepSeek V4.1 Flash | 56.0 |
| GPT-6 Sol | 50.0 |
| Inkling | 44.0 |