Google推出Gemini 2.5电脑使用工具, Google launches Gemini 2.5 Computer Use, an AI that autonomously performs web tasks like shopping and scheduling via text prompts.

Google推出了Gemini 2.5 电脑使用, 这是一种AI模式, Google has launched Gemini 2.5 Computer Use, an AI model that autonomously navigates websites by clicking, typing, and scrolling based on text prompts. 它建在Gemini 2.5 Pro上,利用视觉推理和反馈回路完成复杂的网络任务,如填表、增加购物车和预约。 Built on Gemini 2.5 Pro, it uses visual reasoning and a feedback loop to complete complex web tasks like form filling, shopping cart additions, and appointment scheduling. 该模型支持13个接口行动,在基准方面成绩优异,开发者可通过谷歌AI Studio和Vertex AI向开发者提供,但一般消费者不使用。 The model supports 13 interface actions, excels in benchmarks, and is available to developers via Google AI Studio and Vertex AI, though not for general consumers. 它包括防止风险行动的安全保障,并优化用于基于浏览器的工作流程,而不是全面的系统控制。 It includes safety safeguards to prevent risky actions and is optimized for browser-based workflows, not full system control.