Google has just introduced Gemini 2.5 Computer Use, a new AI model designed to interact directly with a real-life web interface.
powered by the Gemini 2.5 Pro platform, the technology can navigate the browser, fill in forms, roll up pages, click, enter data and use a keyword combination, all through a virtual browser developed by Google.
According to the official blog post, Gemini 2.5 Computer Use has been provided to developers through Google AI Studio and Vertex AI.
The goal of the model is to allow AI to perform complex tasks on the internet based on natural language instructions, such as account registration, data arrangement or software testing.
Google said the model has lower latency and superior performance compared to competitors in many tests of web and mobile standards.
In illustrative videos, Gemini 2.5 Computer Use shows flexible processing capabilities such as AI being able to access websites, read content, and then organize information as requested by users, such as dragging notes into the correct position in a web application.
Google said that these tasks are tripled in speed compared to before, demonstrating the advancement of automatic interface navigation technology.
Currently, Gemini 2.5 Computer Use only supports 13 types of actions, mainly at the browser level and cannot be operated directly at the home computer operating system level.
However, Google confirmed that internal technical teams have applied this model in testing the user interface (UI), significantly shortening the software development time.
This new technology is also integrated in some internal products and projects such as AI Mode in Google Search, Firebase checker and Mariner project (an AI platform that helps users communicate using natural language) to designate autistic agents to perform tasks such as planning, research or data entry.
With Gemini 2.5 Computer Use, Google is taking another step forward in turning AI into a true digital user when it can operate, respond and process information directly on the web, opening up a future where online tasks are fully automated with artificial intelligence.
