Google releases Gemini CLI, an open-source terminal AI agent
Free personal accounts get 60 requests a minute and 1,000 a day against Gemini 2.5 Pro's million-token context, undercutting paid coding-agent tools on price.
- Open weights & ecosystem
- Models & capabilities
- Minor
Google released Gemini CLI, an open-source command-line agent that brings Gemini into developers’ terminals for coding, debugging and task automation. Released under the Apache 2.0 licence, it runs on Gemini 2.5 Pro — the same model behind Google’s Gemini Code Assist extension for VS Code — and inherits its one-million-token context window.
The release’s most consequential detail was pricing rather than capability. Google set the free tier for personal Google accounts at 60 model requests per minute and 1,000 requests per day, quotas described at the time as unusually generous next to competing terminal coding agents that charged per token or capped free usage far lower. Developers wanting higher throughput could pay through Google AI Studio or Vertex AI, or license Gemini Code Assist’s paid tiers.
Gemini CLI arrived into a terminal-agent category that had grown crowded through 2025, alongside tools such as Anthropic’s Claude Code and OpenAI’s Codex CLI, and its open licence and free quotas were read as a deliberate move to win developer mindshare on cost and access rather than on a capability lead. Being open source also meant developers could inspect and modify the tool itself, distinct from the closed-weight model it called. The release added to a broader pattern through the year of frontier labs competing directly for the terminal as a surface for agentic coding, rather than leaving it to third-party wrappers.