Gemini 3.5 Flash Gets Built-In Computer Use
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Gemini 3.5 Flash now has computer use integrated as a native built-in tool, delivering the "best performance yet" for agentic computer use tasks — previously only available as a standalone Gemini 2.5 computer use model.
- Developers can use 3.5 Flash to build custom agents that see, reason, and take action across browser, mobile, and desktop environments, unlocking improved performance for long-horizon tasks like continuous software testing and professional knowledge work.
- Computer use in 3.5 Flash is available via the Gemini API and the Gemini Enterprise Agent Platform, with a demo environment hosted by Browserbase and a reference implementation in the documentation.
- Google applied targeted adversarial training to mitigate prompt injection risks for agents operating in live environments.
- Two optional enterprise safeguards are being released alongside the model: one requires explicit user confirmation for sensitive or irreversible actions, the other automatically stops tasks if an indirect prompt injection is detected.
- Google recommends a "defense-in-depth" approach, urging developers to combine these features with secure sandboxing, human-in-the-loop verification, and strict access controls.
Why it matters: By folding computer use into the main Gemini Flash model rather than keeping it a standalone tool, Google lowers the barrier for enterprises building agents that automate software testing and knowledge work across browsers and apps — but the live-environment risk of prompt injection means the two new safeguard systems (mandatory confirmation, auto-stop on injection) become the gating control for adoption.
