Updated
GPT-6 Luna
As of Oct 20, 2026, GPT-6 Luna ranks #1 of 7 on Bug Bench with a relative bug rate of 0.45 (55% fewer counted bugs than average).
- Relative bug rate
- 0.45
- Rank of 7
- #1
1.00 = average · lower is better
- Provider
- OpenAI
- Released
- Sep 22, 2026
- Price per 1M tokens
- $0.10 in · $0.50 out
- First seen
- Week of Sep 21
Nearby in the ranking
| Rank | Model | Rate |
|---|---|---|
| 1 | GPT-6 Luna | 0.45 |
| 2 | GPT-6.1 Sol | 0.98 |
| 3 | Claude Fable 5.1 | 1.07 |
Is GPT-6 Luna getting worse?
Not enough weekly data yet (4 of 6 weeks).
Where it does better or worse
# = rank among models
By coding agent
| Rank | Agent | Rate |
|---|---|---|
| 1 | Codex | 0.44 |
By language
| Rank | Language | Rate |
|---|---|---|
| 1 | TypeScript | 0.73 |
How widely GPT-6 Luna is used
Share of AI-written code reviewed by cubic
- Latest weekWeek of Oct 12
- 1.4%
- This week so farSince Oct 19
- 0.4%
- Peak weekWeek of Oct 12
- 1.4%
Compare GPT-6 Luna
Questions and answers
- Is GPT-6 Luna good for coding?
- GPT-6 Luna ranks #1 of 7 on Bug Bench, with 55% fewer counted bugs than average (0.45) from Jun 29 to Oct 18, 2026.
- Is GPT-6 Luna getting worse?
- Not enough weekly data yet (4 of 6 weeks).
- Which coding agent has the lowest bug rate with GPT-6 Luna?
- Codex, at 0.44.
Embed GPT-6 Luna's Bug Bench badge
Markdown
[](https://www.cubic.dev/bugbench/models/gpt-6-luna)HTML
<a href="https://www.cubic.dev/bugbench/models/gpt-6-luna"><img src="https://www.cubic.dev/bugbench/badge/gpt-6-luna" alt="GPT-6 Luna on Bug Bench" /></a>