GPT-6.2 Astra and Sol arrive in GitHub Copilot: what the rollout means for US developers
GitHub says OpenAI's GPT-6.2 Astra and Sol are now generally available in Copilot for Pro+, Max, Business and Enterprise plans, rolling out gradually with usage-based billing at provider list prices. OpenAI positions Astra for demanding professional work and Sol for everyday tasks.
What changed
GitHub announced on November 17, 2026 that OpenAI's GPT-6.2 Astra and GPT-6.2 Sol are generally available in GitHub Copilot. According to the changelog, eligible plans are Pro+, Max, Business and Enterprise, and availability is rolling out gradually rather than switching on everywhere at once.
GitHub says model selection is supported in IDEs, the Copilot CLI and on github.com, so users can choose between the two new models in each of those surfaces. The changelog also states that usage-based billing follows provider list pricing, meaning Copilot charges are tied to OpenAI's published rates rather than a separate fixed Copilot price for these models.
- Announced: November 17, 2026 (per GitHub changelog)
- Eligible plans: Pro+, Max, Business, Enterprise
- Where: IDEs, Copilot CLI, github.com
- Billing: usage-based, following provider list pricing
How the two models are positioned
OpenAI describes GPT-6.2 Astra as a model for difficult professional projects and GPT-6.2 Sol as an efficient model for everyday work. That framing suggests a tiered choice rather than a single replacement model: heavier reasoning-oriented work on Astra, routine tasks on the cheaper Sol. Readers should treat the positioning as OpenAI's own description, not an independent benchmark result.
OpenAI reports developer API availability for both models and publishes standard API rates: eight dollars per million input tokens and forty dollars per million output tokens for Astra, and two dollars and ten dollars respectively for Sol. Those figures are OpenAI API list prices. They are not a complete GitHub Copilot bill, and actual Copilot costs will depend on how the billing is applied to a given plan and usage pattern.
- Astra: $8 per million input tokens, $40 per million output tokens (OpenAI standard API rates)
- Sol: $2 per million input tokens, $10 per million output tokens (OpenAI standard API rates)
- These rates are provider list prices, not a full Copilot invoice
What it means for teams and what to watch
For US development teams on Copilot Business or Enterprise, the practical change is a second and third model option inside existing workflows rather than a new tool to adopt. Because GitHub says the rollout is gradual, some users may not see the models immediately, and admins may need to confirm which model choices are enabled for their organization. The changelog excerpt does not detail any separate admin controls, so that should be verified in the product.
The pricing structure is the most consequential detail for budgeting. With usage-based billing at list prices, heavy use of the higher-priced Astra model could raise costs relative to lighter models, while Sol's lower rates make it a plausible default for routine coding assistance. Teams should compare token consumption across models before assuming a flat per-seat cost.
- Confirm model availability in your plan and IDE before standardizing workflows
- Check whether organization-level model policies apply
- Model spend against token usage, especially for Astra
- Track the gradual rollout rather than assuming same-day access