Updated
Best AI model for Go code
As of Nov 14, 2026, Claude Sonnet 5.5 had the lowest relative bug rate in Go code (0.23, 77% fewer counted bugs than the Go average), among 5 ranked models.
| Rank | Model | Rate |
|---|---|---|
| 1 | Claude Sonnet 5.5 | 0.23 |
| 2 | GPT-6 Astra | 0.80 |
| 3= (statistically tied) | Claude Fable 5.1 | 1.26 |
| 4= (statistically tied) | Claude Opus 5.5 | 1.30 |
| 5 | GPT-6.1 Sol | 1.95 |
1.00 = Go average · lower is better · = statistically tied