Google has made Anthropic's Claude Opus 5.5 and Claude Sonnet 5.5 generally available in the Gemini Enterprise app, its AI assistant for companies. The change is listed in the Gemini Enterprise release notes for October 15. Companies on Gemini Enterprise can now offer employees Claude next to Google's own Gemini models, in the same app.
How it works
Off until an admin turns it on. An administrator enables each model in the "Enable third-party models" section of the Feature Management tab. Users then choose Claude in the model selector.
Pay-as-you-go. Claude usage in the app is billed pay-as-you-go.
Regions. The models are available in the global location and in the US and EU multi-regions.
Antigravity got Claude a week earlier
On October 8 Google added the same two models to its AI developer tools: Antigravity 2.0, the Antigravity CLI and Antigravity for IDEs. Here too, third-party models are off by default. Admins opt in through a terms-of-service step in the console, without a separate Anthropic account. Usage is billed as consumption to the company's Cloud Billing account and counts toward the project's spend cap; overages have to be enabled first. Developers can set the thinking level to Low, Medium, High or Max, with Medium as the default. In-country regions are not supported.
Who it matters to
The update is aimed at companies that already run on Google Cloud and Gemini Enterprise. For them, Claude becomes one more model inside a tool their IT team already manages, switched on by an admin and billed by usage, with no separate rollout of another chat app.
For Anthropic it adds another large distribution channel: Claude was already available through Google Cloud for developers, and now it reaches Google's workplace app as well.
What this means if your company uses Gemini Enterprise
If you are an admin, giving people Claude is now a setting rather than a new contract. Because billing follows usage, start with a pilot group and watch spend before you turn it on for everyone.
If you are an employee, ask your admin whether third-party models are enabled. Then run the same real task, such as a report, a long document or a piece of analysis, through Gemini and Claude, and keep the one that does it better.
If your developers work in Antigravity, they can now pick Opus 5.5 or Sonnet 5.5 there. Teams that want Anthropic's own coding agent can compare it with Claude Code.