Google Adds Computer-Use Controls to Gemini 3.5 Flash at Higher Token Price

Google added computer-use controls directly to its Gemini 3.5 Flash model on June 24, enabling the AI to operate desktop, mobile and browser platforms as a human user would, Forbes reported. The update replaces the earlier standalone Gemini 2.5 Computer Use model, which had been limited mainly to browser-based tasks. Gemini 3.5 Flash now controls the physical device in front of the user.

Access is restricted to the Gemini API and the Gemini Enterprise Agent Platform. The feature targets developers and enterprise users for tasks such as interface testing, web research and legacy data entry. Input-token pricing increased to $1.50 per million tokens from $1.25 under the previous model.

Context caching is available to offset some of the added cost. Safety measures include a human-in-the-loop requirement for defined sensitive actions and automatic detection of indirect prompt injection attempts. Developers remain responsible for enabling and using these controls.