Google on Wednesday unveiled Gemini CLI, a free, open-source command-line agent that brings the company’s Gemini 2.5 Pro model directly into developers’ terminals. The tool is available immediately in preview and can be installed with a single npm command. Gemini CLI offers one of the industry’s most generous free allowances—up to 60 model calls a minute and 1,000 a day—while tapping the model’s 1-million-token context window. Released under an Apache 2.0 licence, the software is hosted on GitHub and supports Google Search grounding, the Model Context Protocol for plug-in extensions, and multimodal workflows such as video generation with Veo. The agent shares infrastructure with Gemini Code Assist, allowing users to move between the terminal and IDE-based coding assistants. Individual developers need only a personal Google account to activate the free tier, while enterprise customers can use Vertex AI or AI Studio keys for higher-volume, usage-based billing. By packaging its flagship model in an open-source CLI with an expansive free tier, Google is seeking to woo developers who have gravitated toward rival tools such as OpenAI’s Codex CLI and Anthropic’s Claude Code. The move continues the company’s push to embed Gemini across development workflows after the April release of Gemini 2.5 Pro.