GitHub Adds Claude Opus 5.5 to Copilot With Gradual Access
The new model option is aimed at longer-running coding and knowledge tasks, but GitHub’s efficiency case rests on its own early testing and organization admins retain control over access.
Loading page…
The new model option is aimed at longer-running coding and knowledge tasks, but GitHub’s efficiency case rests on its own early testing and organization admins retain control over access.
Listen to this story
GitHub is expanding Copilot’s model lineup with Claude Opus 5.5 for Pro+, Max, Business, and Enterprise subscribers, but access will be phased in with no universal-availability date. The model will span IDEs, CLI, coding agent, web, and mobile, giving teams one option for longer-running agentic work across their existing workflows. GitHub’s evidence is internal early testing: comparable task resolution to Claude Opus 5 with fewer steps and tokens. Enterprise administrators can control access, while usage followsတာ?
The rollout covers VS Code, Visual Studio, JetBrains, Xcode, Eclipse, Copilot CLI, coding agent, web, and mobile.
GitHub reports comparable task resolution to Claude Opus 5 using significantly fewer steps and tokens in early tests.
Eligible users may not see the model immediately; GitHub provides no date for universal availability.
GitHub is adding Claude Opus 5.5 to GitHub Copilot with a promise of more efficient work on multistep tasks. But the evidence behind that pitch is GitHub’s early testing, and the model will arrive gradually rather than appearing for every eligible user at once.
The rollout gives Copilot Pro+, Max, Business, and Enterprise subscribers another selectable model across editors, command-line tools, GitHub’s coding agent, the Copilot app, and mobile. GitHub says Claude Opus 5.5 can handle agentic coding, long-running agentic tasks, and knowledge work.
GitHub says its early tests found Opus 5.5 resolved tasks comparably to Claude Opus 5 while using significantly fewer steps and tokens. It also says the model recovered quickly from errors in multistep work. Those results describe GitHub’s testing, rather than an independent comparison.
The same model choice is scheduled to appear in Visual Studio Code, Visual Studio, Copilot CLI, GitHub Copilot coding agent, the Copilot app on github.com, GitHub Mobile, JetBrains IDEs, Xcode, and Eclipse. That range puts an option designed for lengthy agent work beside ordinary interactive coding tools, rather than confining it to a separate service.
There is no promised date for universal availability. GitHub says the deployment is gradual and tells users who cannot yet see the model to check back later. For teams, that means the announced model choice may not be simultaneously available to every developer.
Business and Enterprise administrators can manage access to Opus 5.5 through Copilot’s model policies. Under GitHub’s default model-enablement setting, new models are automatically enabled unless an administrator has turned off the global default or explicitly disabled this one.
Usage is billed at provider list pricing under Copilot’s usage-based billing. GitHub does not list a separate Copilot price for Opus 5.5 in the announcement.
GitHub says Claude Opus 5.5 watermarks its text outputs. According to GitHub, the watermark does not alter an output’s meaning, quality, or readability, and it does not add tokens or cost.
Loading discussion...
Join the conversation
Explain what should trigger an administrator’s review.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.