Updated

GPT-6 Astra

As of Nov 9, 2026, GPT-6 Astra ranks #4 of 7 on Bug Bench with a relative bug rate of 0.95 (5% fewer counted bugs than average), statistically tied with Claude Fable 5.1 and Claude Sonnet 5.5.

Relative bug rate
0.95
Rank of 7
#4
Tied with
Claude Fable 5.1 and Claude Sonnet 5.5

1.00 = average · lower is better

Provider
OpenAI
Released
Sep 4, 2026
Price per 1M tokens
$10 in · $50 out
First seen
Week of Aug 31

Nearby in the ranking

GPT-6 Astra and the models ranked closest to it, by relative bug rate.
RankModelRate
2GPT-6.1 Sol
0.74
3= (statistically tied)Claude Fable 5.1
0.88
4GPT-6 Astra
0.95
5= (statistically tied)Claude Sonnet 5.5
1.04
6Claude Opus 5.5
1.11

Is GPT-6 Astra getting worse?

Not enough weekly data yet (0 of 6 weeks).

Where it does better or worse

# = rank among models

By coding agent

GPT-6 Astra relative bug rate by coding agent.
RankAgentRate
3Codex
0.96

By language

GPT-6 Astra relative bug rate by language.
RankLanguageRate
6TypeScript
1.08
5JavaScript
1.63
5Python
1.18
2Go
0.88
2Rust
0.36
3Swift
0.90
2C#
0.60
3Ruby
1.25
5SQL
0.94
1Shell
0.67
2Vue
0.69
2Svelte
2.24
1Terraform
0.44

How widely GPT-6 Astra is used

Share of AI-written code reviewed by cubic
Latest weekWeek of Nov 2
4.5%
Peak weekWeek of Sep 7
22.7%

Compare GPT-6 Astra

Questions and answers

Is GPT-6 Astra good for coding?
GPT-6 Astra ranks #4 of 7 on Bug Bench, with 5% fewer counted bugs than average (0.95) from Jul 20 to Nov 8, 2026.
Is GPT-6 Astra getting worse?
Not enough weekly data yet (0 of 6 weeks).
Which coding agent has the lowest bug rate with GPT-6 Astra?
Codex, at 0.96.
Embed GPT-6 Astra's Bug Bench badge

Markdown

[![GPT-6 Astra on Bug Bench](https://www.cubic.dev/bugbench/badge/gpt-6-astra)](https://www.cubic.dev/bugbench/models/gpt-6-astra)

HTML

<a href="https://www.cubic.dev/bugbench/models/gpt-6-astra"><img src="https://www.cubic.dev/bugbench/badge/gpt-6-astra" alt="GPT-6 Astra on Bug Bench" /></a>