On June 24, 2026, Google announced that computer use is now a native, built-in capability of Gemini 3.5 Flash, moving it out of a separate standalone model and into the main Flash offering. The tool lets agents see, reason, and take action across browser, mobile, and desktop environments for tasks such as continuous software testing and knowledge-work automation.
Google made the capability generally available, accessible immediately through the Gemini API and the Gemini Enterprise Agent Platform, with a demo environment via Browserbase and a GitHub reference implementation. The announcement references OSWorld benchmark results showing Flash’s performance, though specific numbers were not detailed in the post body.
On safety, Google cited targeted adversarial training plus optional enterprise safeguards, including user-confirmation requirements for sensitive actions and automatic task stopping when prompt injection is detected.