Google has officially launched Gemini CLI, a powerful command-line interface (CLI) that brings its Gemini 2.5 Pro AI agent directly into developers’ terminals on Windows, macOS, and Linux. Released on June 25, 2025, the open-source tool (Apache 2.0‑licensed) empowers users to perform natural-language–driven coding, debugging, research, content generation, and even image/video creation via Veo and Imagen integrations.

Built atop the Gemini 2.5 Pro model, Gemini CLI supports an enormous 1 million-token context window, enabling seamless interaction with large codebases or complex tasks—all within the terminal. The free preview offers an unmatched allotment of 60 model requests per minute, and 1,000 daily, with Google touting this as the “largest allowance” in the industry.

Integration features include support for the Model Context Protocol (MCP), Gemini Code Assist compatibility, real-time Google Search grounding, and seamless scripting via Node‑based installation (npx …)—making it a versatile tool for developers who prefer terminal workflows. Users on social media have welcomed Gemini CLI enthusiastically, calling its free tier “game‑changing” and praising the million‑token window as a standout innovation. However, some voices on Reddit mention quirkiness in token limit implementation and confusion over Workspace account eligibility.

Gemini CLI directly challenges existing developer AI solutions like Anthropic’s Claude Code and OpenAI’s Codex CLI—not just in performance, but in its open‑source nature and aggressive access policy. As competition escalates, Google’s offering could reshape AI tooling adoption, driving broader demand for terminal‑first developer experiences.