title: "Google ships native computer-use to Gemini 3.5 Flash" slug: "google-ships-native-computer-use-to-gemini-35-flash" published: "2026-07-25" beat: "Launches" tags: ["Launches", "Tools"] creator: "Agentry Newsroom" editor: "Susanne Sperling, Editor — Human in the Loop" tools: ["Claude (Anthropic)", "Perplexity Sonar"] creativeWorkStatus: "verified" dateReviewed: "2026-07-25" aiActArticle50: "compliant" humanView: "https://agentry.news/google-ships-native-computer-use-to-gemini-35-flash" agentView: "https://agentry.news/agent/google-ships-native-computer-use-to-gemini-35-flash"
Google integrated native computer-use capability into Gemini 3.5 Flash on June 24, 2026, enabling developers to build custom agents that can see, reason, and take action across browser, mobile, and de
Drafted by an AI agent. Verified by Susanne Sperling, Editor — Human in the Loop. AI policy.
Google shipped native computer-use capability for Gemini 3.5 Flash on June 24, 2026, a built-in feature that enables developers to build custom agents capable of perceiving and acting across browser, mobile, and desktop interfaces Google.
The feature integrates into the Gemini API, allowing agents to see screen content, reason about user intent, and execute actions without requiring external tool bindings or middleware. Developers can now deploy agentic systems directly through Google's Enterprise Agent platform with computer-use baked into the model itself.
With native computer-use support, Gemini 3.5 Flash agents can navigate browsers, control mobile interfaces, and interact with desktop applications by processing visual input and executing programmatic actions. This eliminates the need for separate orchestration layers or custom vision-and-action pipelines, compressing the developer workflow from design through deployment Google DeepMind.
The capability applies to a range of use cases: customer service automation, form-filling, application testing, workflow automation, and data extraction tasks that require agents to understand context visually before acting. By moving computer-use into the model itself, Google positions Gemini 3.5 Flash as a direct competitor to Claude's computer-use offering, which has been available through Anthropic's API since earlier in 2026.
The feature is available to developers via the Gemini API and the Enterprise Agent platform. Google has not announced usage-based pricing changes specific to computer-use transactions, though API costs follow standard Gemini 3.5 Flash pricing. Third-party documentation from developer platforms confirms the rollout is active, though some technical documentation still bears update timestamps from earlier in June Android Authority.
This launch closes a technical gap between Gemini and Claude in the agent-building space. Anthropic's Claude model family added native computer-use in early 2026, and this Google update signals a competitive acceleration in bundling vision, reasoning, and action into foundation models rather than requiring developers to bolt components together.
The timing aligns with enterprise adoption cycles—mid-year feature releases often precede Q3 budget deployments and procurement decisions. Google's inclusion of computer-use in both the standard API and enterprise offerings suggests confidence in the feature's stability and developer demand.