Google Launches ‘Gemini 2.5 Computer Use’: AI That Can Click, Type, and Scroll Like Humans | Key Features

Google has introduced Gemini 2.5 Computer Use, a new AI model that can interact with web browsers just like a human. Unlike traditional systems that depend on structured APIs, this AI uses visual understanding to carry out tasks such as clicking buttons, filling out forms, scrolling pages, and navigating websites.

It is designed to work smoothly across interfaces that humans typically use, making browser automation more natural and intuitive.

What Is Gemini 2.5 Computer Use?

Gemini 2.5 Computer Use is built on the Gemini 2.5 Pro platform, which provides advanced visual understanding and reasoning skills. It powers AI agents capable of interacting with user interfaces (UIs) and has been shown to outperform leading alternatives on several web and mobile control benchmarks.

Additionally, it offers lower latency, making it faster and more efficient than many other models.

ALSO READ: Indian Creator Turns Elon Musk Into Iron Man Using Grok Imagine; Musk Responds | Watch

Gemini 2.5: Key Features

• Human-Like Interaction: The model can carry out 13 predefined actions, such as clicking, typing, scrolling, hovering, and using keyboard shortcuts. These capabilities let it mimic human behaviour in web browsing accurately.
• Developer Access: Developers can access the model through Google AI Studio and Vertex AI, enabling integration into various apps and workflows.
• Benchmark Performance: Gemini 2.5 Computer Use beats competing models in tests and shows lower latency, making it a reliable choice for browser automation tasks.

Gemini 2.5: Practical Applications

The model is especially useful for tasks that need direct interaction with web pages, such as UI testing, automating workflows, or handling web tasks without a direct API. Its ability to understand the layout and elements of web pages makes it a powerful tool for developers, allowing them to automate complex browser-based tasks efficiently.

ALSO READ: Africa Is Splitting: What the East African Rift Means for Earth’s Future

Gemini 2.5: Limitations

Currently, Gemini 2.5 Computer Use works only in web browsers. It is not yet optimized for desktop operating systems. The model supports only 13 predefined actions, which limits its capabilities outside browser environments. For now, it is primarily focused on browser-based tasks.

Gemini 2.5: Safety Measures

Google has put in place safety controls to reduce risks related to AI misuse. These include preventing harmful actions, stopping prompt injections, and avoiding scams within the browser. Developers can use these controls to ensure the AI does not automatically perform risky tasks, keeping its use safer for testing and automation.

ALSO READ: Meet Vyommitra: India’s Spacefaring Humanoid Set to Revolutionize Gaganyaan Mission

What’s Next for Gemini 2.5?

Gemini 2.5 Computer Use marks a major step forward in AI that can interact with computers like humans. Its capabilities open the door to new ways of automating complex browser-based tasks, improving efficiency for software testing and workflow automation.

As Google continues to refine the model, it may eventually expand beyond browsers to handle more desktop-level tasks and broader automation challenges.