CWE-bench v1: Go
Code · 2026-09-25
Programmatic pass@1 across the 26 Go tasks in CWE-bench v1, averaging four rollouts per task.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 4 Argon | 76.0 |
| Claude Opus 5.5 | 68.0 |
| Grok 4.7 | 65.0 |
| GPT-6 Astra | 62.0 |
| Hy4 Preview | 62.0 |
| DeepSeek V4.1 Flash | 61.0 |
| Grok 4.6 | 60.0 |
| Muse Spark 1.3 | 59.0 |
| Claude Fable 5.1 | 51.0 |
| GPT-6 Sol | 43.0 |
| Inkling | 43.0 |