GPT-6.1 Sol arrives in GitHub Copilot
OpenAI's GPT-6.1 Sol is now generally available and rolling out in GitHub Copilot. The model targets agentic coding and terminal workflows, with early testing showing it completes tasks using fewer tokens and steps than earlier models in the GPT-6 and GPT-5.6 families.
Copilot Pro+, Max, Business, and Enterprise users can select GPT-6.1 Sol from the model picker. Availability spans Visual Studio Code, Visual Studio, Copilot CLI, the GitHub Copilot coding agent, the GitHub Copilot app, github.com, GitHub Mobile on iOS and Android, JetBrains IDEs, Xcode, and Eclipse. Rollout is gradual, so the option may not appear immediately for every user.
What early testing shows
GitHub's testing found the model "reliably completed tasks while using noticeably fewer tokens and steps" than its predecessors. That efficiency matters for teams watching usage-based billing costs. The model is billed at provider list pricing under Copilot's usage-based billing structure.
For developers, fewer tokens per task translates directly to lower cost per completed workflow. The reduction in steps also suggests less back-and-forth during multi-step coding and terminal operations.
Administrator controls
Copilot Business and Enterprise administrators can manage access through the model policy in Copilot settings. Under default model enablement, new models are automatically enabled unless an administrator has turned off the global default or explicitly disables GPT-6.1 Sol.
Organizations with strict model governance policies can review their settings before the model reaches their developers. The model picker gives individual users visibility into what's available on their plan.
Why this matters for IT and development teams
Token efficiency isn't just a technical detail. For teams running Copilot at scale, a model that completes the same work with fewer tokens reduces per-seat costs under usage-based billing. IT administrators should review their model policies now, before rollout completes, to decide whether GPT-6.1 Sol should be available to their developers on day one.
Developers working with AI Code Generation Courses or evaluating Copilot's coding agent should test the model against their own workflows. The early efficiency claims are worth validating on real repositories, not just benchmarks.
Your membership also unlocks:
