PANews reported on June 25 that according to The Verge , Google has released a new open source AI agent tool , Gemini CLI , which brings Gemini 's programming, content generation and research capabilities directly to developers' terminals. Gemini CLI is based on the Gemini 2.5 Pro inference model, supports 1 million token context windows, and integrates Gemini Code Assist , MCP protocol and Google search. It can also call Veo and Imagen tools to generate pictures and videos. Currently, the tool can be applied for free through a personal Google account, with a usage limit of 60 times per minute and 1,000 times per day. Google claims this is the largest free quota in the industry.