First-party only — a session id and where you arrived from, so we can tell which pages are worth writing more of. No advertising, no third-party trackers, nothing that follows you off this site.
Both are on this list because they are genuinely good at something. The question is which something you need. Below is the factual comparison, then the honest answer about who each one is for.
Reviewed July 25, 2026 · pricing changes often, check the vendor
| Gemini CLI | Claude Code | |
|---|---|---|
| Made by | Anthropic | |
| Where it runs | Terminal | Terminal |
| Models | Gemini models | Claude models only (Opus / Sonnet / Haiku tiers) |
| Pricing | Free tier with a personal Google account · paid for higher limits | From $20/mo (Pro) · usage-based API also available |
| Open source | Yes | No |
| Best for | Trying a terminal agent seriously without a subscription, or working with very large contexts. | Multi-file work in a repo you can describe well — the more context you put in CLAUDE.md, the further ahead it gets. |
The free tier is real and large enough to do actual work
Very large context windows suit whole-repo questions
Apache-2.0 licensed, so you can read and fork it
Gemini models only
Younger than the alternatives; conventions are still moving
Free tier terms are subject to change
Configuration is a first-class feature — memory files, skills, hooks and permissions are documented and versionable in the repo
Works on the real filesystem and the real shell, so it can run your tests instead of guessing whether they pass
Subagents let you fan work out without polluting the main context window
Anthropic models only — you cannot point it at a different provider
Terminal-first, which is a poor fit if you want inline autocomplete while you type
Heavy usage on the $20 plan runs into limits fast; the useful tiers cost more
Pick Gemini CLI if trying a terminal agent seriously without a subscription, or working with very large contexts.
Pick Claude Code if multi-file work in a repo you can describe well — the more context you put in CLAUDE.md, the further ahead it gets.
And a thing worth saying plainly: the gap between these two is smaller than the gap between a repo that tells an agent what it is and one that does not. If you are choosing between them to fix inconsistent output, the tool is probably not the variable. The free Foundations track covers the part that is.