GPT-6 Astra Goes Generally Available in GitHub Copilot
GitHub is rolling the OpenAI model out to Pro+, Max, Business and Enterprise seats, billed at provider list pricing. No benchmark numbers given.
Symbolic image: a developer workstation at night, three monitors with abstract code blocks, a small machine with blinking status lights beside the desk while an automated run proceeds.
OpenAI's GPT-6 Astra became generally available in GitHub Copilot on September 4, 2026, across four plans: Pro+, Max, Business and Enterprise.
At a glance
- Announced in the GitHub changelog on September 4, 2026.
- Available to Copilot Pro+, Max, Business and Enterprise; the entry names neither Free nor Pro.
- Ten surfaces: VS Code, Visual Studio, Copilot CLI, coding agent, Copilot app, github.com, GitHub Mobile, JetBrains, Xcode, Eclipse.
- Billed at provider list pricing under usage-based billing; GitHub states no premium-request multiplier.
- Rollout is gradual; Business and Enterprise admins control access via model policy in Copilot settings.
OpenAI's GPT-6 Astra is generally available in GitHub Copilot as of September 4, 2026. The changelog names four plans: Copilot Pro+, Max, Business and Enterprise. Plain Copilot Pro and Copilot Free are absent from that list. The rollout is gradual, and GitHub tells anyone who cannot see the model yet to check back later.
Where the model shows up
GitHub lists ten surfaces where Astra can be picked: Visual Studio Code, Visual Studio, the Copilot CLI, the Copilot coding agent, the Copilot app, github.com, GitHub Mobile on iOS and Android, JetBrains IDEs, Xcode and Eclipse. That covers both the interactive path inside an editor and the unattended path where Copilot works through an issue on its own. The entry gives no order in which those surfaces light up.
What GitHub claims about it
GitHub describes Astra as built for long-horizon, autonomous coding work. In its telling, the model plans and validates as it goes, batches diagnosis with verification, confirms its own results, and reaches stronger performance on long tasks in fewer steps than earlier OpenAI models.
All of that is vendor language published in a vendor channel. The entry carries no benchmark, no percentage and no head-to-head comparison against another model. A reader who wants to test the claim gets nothing from this source to test it against.
Cost and the admin switch
On billing, GitHub says the model is charged at provider list pricing under usage-based billing, and links its pricing documentation. The premium-request multiplier that GitHub publishes for some other models does not appear in this text.
For organizations, the relevant control is model policy in Copilot settings. Under default model enablement, new models turn on automatically unless an administrator switches off the global default or disables this specific model. Teams under procurement or compliance constraints will want to check that before the rollout reaches them rather than after.
What stays unverified
This report rests on the vendor announcement alone. As of publication, no independent newsroom had confirmed the rollout or tested the performance claims.
That leaves unverified: any figure for speed or accuracy, a possible premium-request multiplier, the date the gradual rollout completes, and whether any older models are being retired to make room. The entry itself names no model that Astra replaces.
FAQ
Which Copilot plans include GPT-6 Astra?
Per the changelog, Copilot Pro+, Max, Business and Enterprise. Copilot Free and Copilot Pro are not named in the entry.
How is GPT-6 Astra billed in Copilot?
GitHub says it is charged at provider list pricing under usage-based billing and links its pricing docs. No premium-request multiplier is stated in the entry.
Can an administrator turn GPT-6 Astra off?
Yes, through model policy in Copilot settings. Under default model enablement new models are on automatically until the global default or this model is explicitly disabled.
