This page compares the agents. For the models behind them, see Codex vs Claude for coding. Claude Code and Codex CLI are both terminal agents that read your repository, edit files and run commands, and both also come as IDE, desktop and web versions. Claude Code runs Anthropic's models: on Pro, Max, Team, Enterprise and API accounts its default is Claude Opus 5.5, and you can switch to Claude Fable 5.1 with /model fable. Codex CLI runs OpenAI's models, and you choose the model and reasoning effort with /model, including GPT-6 Astra and GPT-6.1 Sol.
Day to day, the two share a lot: project instruction files (Claude Code reads CLAUDE.md and can also read the AGENTS.md file Codex uses), skills, MCP servers, subagents, and a non-interactive mode for scripts and CI (claude -p and codex exec). The differences show up at the edges. Codex CLI is open source under the Apache 2.0 licence, has a built-in /review command for local code review and can hand tasks to Codex cloud. Claude Code adds hooks that run shell commands around its actions, scheduled routines, and GitHub Actions and GitLab CI/CD integrations. Access follows the subscription: Codex signs in with ChatGPT, and Claude Code needs a Claude plan or an Anthropic Console account.
Which should you choose?
Claude Code
Claude Fable 5.1 · 1M context
Pick it if
you want Claude models, hooks around every action and scheduled routines.
Codex CLI
GPT-6 Astra · 1M context
Pick it if
you want OpenAI models, an open-source CLI, /review and a hand-off to Codex cloud.
Both agents cover the same core loop and share conventions like AGENTS.md, skills and MCP, so the model you prefer usually decides. Pick Claude Code for Anthropic's models, hooks and scheduled routines. Pick Codex CLI for OpenAI's models, an open-source CLI and a built-in review command.
Questions
Which models do Claude Code and Codex CLI use?
Claude Code defaults to Claude Opus 5.5 on Pro, Max, Team, Enterprise and API accounts, and Claude Fable 5.1 is available with /model fable. On some plans Fable usage bills to usage credits. Codex CLI lets you choose among OpenAI models with /model, including GPT-6 Astra and GPT-6.1 Sol.
Can I use both on the same repository?
Yes. Claude Code can read the AGENTS.md file that Codex uses, so one set of project instructions can serve both, and both support MCP servers for shared tools.
Do they run in CI?
Yes. Codex has codex exec for non-interactive runs, and Claude Code has claude -p plus GitHub Actions and GitLab CI/CD integrations.