Tech Desk 
Oct 08, 2025
As artificial intelligence (AI) seems to be a toned-up version of automation, the latter seems to have been coupled with the former, as Google has introduced the Gemini 2.5 Computer Use AI, its advanced AI model capable of mirroring human interaction to carry out complex digital tasks.
The Gemini 2.5 Computer Use AI can perform on-screen actions like clicking, typing, and scrolling, making it a subtle replacement of human computer operators.
Built on the Gemini 2.5 Pro framework, the Gemini 2.5 Computer Use AI is equipped with impressive visual reasoning and navigation skills to operate across web browsers and Android systems. This is an impressive advancement in AI technology aimed at automating computer workflows efficiently.
Google claimed that Gemini 2.5 excels in various benchmarks and outperforms competitors. In the WebVoyager test, it scored 88.9%, beating OpenAI’s Computer-Using AI Agent, which achieved 87%. It also performed well on the Online-Mind2Web benchmark, outpacing both OpenAI’s and Anthropic’s Claude Sonnet 4.5 models.
Google said: “Gemini 2.5 Computer Use is a step toward creating more capable AI agents that can operate independently and assist users across digital platforms.”
Google confirmed that Project Mariner and the AI Mode in Google Search are already using different versions of the Gemini 2.5 Computer Use AI.
The developer-facilitating end of the Gemini 2.5 Computer Use AI is that developers can now access its API through Google AI Studio and Vertex AI, to test its applications in various industries.
