Google’s New AI Performs Human-Like Computer Tasks

Google’s Gemini 2.5 Pro Outperforms OpenAI in Computer Use AI Benchmark.Google has launched a new AI model called Gemini 2.5 Computer Use. This advanced system is designed to operate computers by controlling user interfaces. It is built upon the powerful Gemini 2.5 Pro architecture.The model can navigate web browsers and Android applications. It performs tasks by clicking, typing, and scrolling autonomously. This development marks a significant step towards creating practical AI agents.

Benchmark Results Show Superior Performance

Google’s new model has demonstrated leading capabilities in industry tests. According to Reuters, it achieved a score of 88.9% on the WebVoyager benchmark. This slightly outperforms OpenAI’s competing computer-using agent.In the Online-Mind2Web benchmark, Google also secured a higher score. The model shows improved accuracy and lower latency compared to rivals. This includes Claude Sonnet 4.5 and the OpenAI operator.These results indicate a robust AI for real-world tasks. The technology can reliably complete complex assignments on digital interfaces. This could significantly boost productivity and automation.

Practical Deployment and Future Applications

Google is already implementing this technology in its own products. Versions of the model are active in Project Mariner and AI Mode for Google Search. This provides a real-world testing ground for its capabilities.The API is now accessible to developers through Google AI Studio and Vertex AI. This allows for broader experimentation and integration. The potential applications span from customer service to software testing.The race for superior AI agents is clearly intensifying. Google’s latest release positions it as a frontrunner in this critical field. The impact on how we interact with software could be profound.

Commonwealth Champ Sangram Singh’s European MMA Debut

This advancement in computer-use AI represents a key milestone. The Gemini 2.5 Pro model sets a new benchmark for practical AI agents. Its real-world integration is just beginning.

Info at your fingertips

What is Gemini 2.5 Computer Use?

It is a new AI model from Google built on the Gemini 2.5 Pro. It is specifically trained to operate computer interfaces like browsers and Android apps. The model can perform clicks, typing, and scrolling to complete tasks.

How does it compare to OpenAI’s model?

Google’s model slightly outperformed OpenAI’s agent in key benchmarks. It scored 88.9% versus 87% in the WebVoyager test. It also showed better results in the Online-Mind2Web evaluation.

Where is this technology being used now?

Google has already deployed it within Project Mariner and its AI Search mode. Developers can also access the API through Google AI Studio and Vertex AI. This allows for building custom applications.

What is the main advantage of this AI?

The main advantage is its ability to reliably automate complex digital tasks. It demonstrates high accuracy and low latency in tests. This makes it a powerful tool for productivity and software interaction.

Why is computer-use AI important?

It is a critical step towards creating general-purpose AI agents. These systems can interact with software just like a human would. This could revolutionize automation across many industries.