Gemini 3.5 Flash gets built-in computer use capability

Google DeepMind has introduced native support for the "computer use" function in the Gemini 3.5 Flash model, which was previously only available as a separate Gemini 2.5 computer use model. Developers can now use 3.5 Flash to create agents capable of performing tasks with long-horizon planning and automating enterprise processes, such as continuous software testing and working with professional applications. Access to the function is provided via the Gemini API and the Gemini Enterprise Agent Platform. To reduce the risks of prompt injection, DeepMind applied targeted adversarial training and released two optional enterprise protection systems: requiring explicit user confirmation for sensitive actions and automatically stopping tasks upon detecting indirect prompt injection. The company recommends combining these measures with secure isolation, human-in-the-loop verification, and strict access control. Users are invited to test the capabilities in the Browserbase demo environment.