GitHub Adds Claude Sonnet 5.5 Across Copilot
The Anthropic model is entering Copilot's editors, command line, coding agent, app and mobile clients under a gradual rollout with enterprise policy controls.
Edited by Tyronne Panaino
GitHub announced on September 28 that Claude Sonnet 5.5 is generally available in GitHub Copilot for Pro, Pro+, Max, Business and Enterprise users. The rollout spans desktop editors, the command line, GitHub's coding agent and app, github.com, mobile clients and several third-party IDEs, giving paid users another model choice for everyday feature work and bug fixing.
The change matters because it is not confined to one editor integration. GitHub is placing the model behind the same Copilot selection and enterprise-policy layer across a wide range of work surfaces, while warning that availability is gradual and may not appear for every eligible user immediately.
One model across many Copilot surfaces
GitHub lists Visual Studio Code, Visual Studio, Copilot CLI, the Copilot coding agent, the GitHub Copilot app and github.com among the supported surfaces. GitHub Mobile on iOS and Android is included, as are JetBrains IDEs, Xcode and Eclipse. Users select Claude Sonnet 5.5 through the relevant model picker when the rollout reaches their account.
That breadth can reduce the gap between an organization's approved model list and the places developers actually invoke Copilot. A team could, for example, encounter the same model option in an IDE, a terminal workflow and a delegated coding-agent task. The announcement establishes availability across those surfaces, but it does not say that outputs, tool access or task behaviour are identical in every client. Teams should therefore validate the model separately in the environments and workflows they depend on.
GitHub reports efficiency gains, not an independent benchmark
GitHub describes Claude Sonnet 5.5 as intended for well-scoped everyday work such as building features and fixing bugs. In GitHub's early testing, the company says the model matched Claude Sonnet 5 on coding tasks while using fewer steps, tokens and tool calls, and completed tasks faster.
Those are first-party observations rather than independently reproduced results. The changelog does not publish the evaluation set, sample size, exact token reduction, latency measurements or error analysis. It also does not establish that fewer tool calls will improve every repository or agent workflow. The practical checkpoint is whether teams see the same efficiency pattern on their own codebases without losing correctness, reviewability or policy compliance.
Administrators retain model-policy control
Copilot Business and Enterprise administrators can manage access through the model policy in Copilot settings. Under GitHub's default model-enablement behaviour, new models are enabled automatically unless an administrator has disabled the global default or explicitly blocked this model. Organizations with a restricted model catalogue should review that setting rather than assume the new option is either universally available or universally disabled.
GitHub says usage is billed at the provider's list pricing under usage-based billing. The announcement does not state a price in the changelog itself, so this article does not infer a cost or compare the model with alternatives on price. Plan eligibility, gradual availability and enterprise policy are the verified release facts.
Status
Confirmed. GitHub has documented general availability, supported Copilot plans and clients, gradual rollout, billing treatment and administrator controls. Internal confidence is medium because the availability and efficiency claims come from one first-party announcement and were not independently tested in this run.
Sources
Update note: Last reviewed 2026-09-28. We will revise this post if GitHub changes rollout scope, plan access, pricing treatment or model-policy controls.
Sources
Drafted with AI assistance from source briefs; reviewed for citation completeness and label accuracy.