Claude Code vs Codex: How They Actually Differ
Codex fits builders who want one agent that moves across CLI, IDE, desktop, web, and cloud, with verified pricing starting at free. Claude Code fits terminal-first developers who want Anthropic model depth and richer workflow primitives like skills, hooks, and slash commands.
Compiled 2026-08-21 from our catalogue entries for Claude Code and Codex.
The short answer
- Pick Claude Code if
- You live in the terminal, want Anthropic Claude models (Opus, Sonnet, Haiku), and value native MCP as both client and server along with skills, hooks, and slash commands for encoding your own workflow.
- Pick Codex if
- You want a single agent that works across CLI, IDE, desktop, web, and cloud, want GitHub pull request review that proposes fixes, and want verified pricing tiers included with ChatGPT plans from the free tier up.
- Cheapest way in
- Codex has a Free plan at $0 per month. Claude Code's starting price is not recorded on file, so the cheapest confirmed entry point between the two is Codex Free.
- What we could not compare
- Claude Code's specific plan tiers and starting price are not on file. Public repos and docs are not listed for either product, and Business / Enterprise Codex pricing is not published.
Claude Code is Anthropic's official command-line agent. It runs in the terminal on macOS, Windows, and Linux, edits files, executes commands, and ships pull requests, and is built around first-class support for Opus, Sonnet, and Haiku. It is freemium, but specific plan tiers and a starting price are not on file, and there is no IDE or web surface listed.
Codex is OpenAI's coding agent. The same agent works across CLI, IDE extension, desktop, web, and cloud, with verified ChatGPT plan tiers from Free ($0/month) up to Pro 20x ($200/month), plus a pay-per-token API key option. Both tools support MCP and read the shared AGENTS.md standard, but Claude Code is locked to Anthropic models and Codex is locked to OpenAI's GPT models.
Side by side
Every cell is a field on our catalogue entry for that tool, copied as recorded. Rows where the two differ come first; blank fields say so rather than being filled in.
| Claude Code | Codex | |
|---|---|---|
| Starts at | Not recorded | Free — $0 /month |
| Runs on | mac, win, linux, cli | web, mac, win, linux, cli, vscode-ext |
| Category | AI Coding Agents | AI Coding Agents |
| Pricing model | Freemium | Freemium |
| Public source code | No public repo listed | No public repo listed |
| Public docs | Not listed | Not listed |
- Public source code: From the repository link on each catalogue entry, not a licence audit.
Where Claude Code is the better fit
- First-class support for Anthropic Claude Opus, Sonnet, and Haiku.
- Native MCP server protocol support as both client and server.
- Skills, hooks, and slash commands let teams encode their own workflow primitives.
Where Codex is the better fit
- One agent across CLI, IDE extension, desktop, web, and cloud, with work moving between surfaces.
- GitHub integration with cloud code review on a pull request that catches bugs and proposes fixes.
- Reads the shared AGENTS.md standard, supports MCP tools, and is included with ChatGPT plans from the free tier up.
Which one should you pick?
It is included with ChatGPT plans starting at the free tier, and the cloud code review on GitHub pull requests is a verified capability on the fact sheet.
It is built around Opus, Sonnet, and Haiku with native MCP as both client and server, and ships skills, hooks, and slash commands for custom primitives.
Its verified surfaces include web, macOS, Windows, Linux, CLI, and a VS Code extension, with a free tier entry and a published plan ladder up to Pro.
Frequently asked questions
Frequently asked questions
- Can Claude Code run anywhere besides the terminal?
- The fact sheet lists platforms as mac, win, linux, and cli. No IDE or web surface is listed for Claude Code.
- Do I need a paid ChatGPT subscription to use Codex?
- No. The verified plans include a Free $0/month tier, and the description states Codex is included with ChatGPT plans from the free tier up through Pro.
- Can either tool use the other company's models?
- No. Claude Code is locked to Anthropic models, and Codex is locked to OpenAI's GPT models, with no Claude or Gemini support out of the box for Codex.
- Are docs or public repos available for either product?
- Public repos and docs are listed as none listed for both Claude Code and Codex on the fact sheet.
What we compared
This page states nothing that is not in the two catalogue entries below. The exact text used to write it is reproduced here so you can check our reading, and both tools link through to their own pages and to the vendor.
- /tools/claude-code — checked 2026-05-25 · official site
- /tools/codex — checked 2026-07-02 · official site
Show the source text
## Claude Code (/tools/claude-code) Tagline: Anthropic's official CLI agent for Claude — runs in the terminal, edits files, executes commands, and ships PRs. Category: AI Coding Agents Pricing model: freemium Starts at: Not recorded Platforms: mac, win, linux, cli Public repo: none listed Docs: none listed Verified plans: none on file Strengths on file: First-class support for Anthropic Claude Opus, Sonnet, and Haiku.; MCP server protocol is natively supported as both client and server.; Skills, hooks, and slash commands let teams encode workflow primitives. Limits on file: Locked to Anthropic models — no GPT or Gemini support out of the box.; Terminal-first UX has a learning curve for developers expecting an IDE. Claude Code is Anthropic's official command-line agent for Claude — it runs in the terminal, edits files, executes commands, and ships pull requests. Its strength is depth on Claude and extensibility: first-class support for Opus, Sonnet, and Haiku, native MCP as both client and server, and skills, hooks, and slash commands that let teams encode their own workflow primitives. It's for developers comfortable in the terminal who want an agent that acts, not just autocompletes. Two trade-offs: it's locked to Anthropic models — no GPT or Gemini out of the box — and the terminal-first UX has a learning curve if you're expecting an IDE. In the stack it's the terminal-first agent next to IDE-native Cursor and the open, model-agnostic Aider. ## Codex (/tools/codex) Tagline: OpenAI's coding agent — works in the terminal, IDE, and cloud, reviews GitHub pull requests, and runs on GPT-5 models. Category: AI Coding Agents Pricing model: freemium Starts at: Free — $0 /month Platforms: web, mac, win, linux, cli, vscode-ext Public repo: none listed Docs: none listed Verified plans (read 2026-08-21): Free $0 /month | Go $8 /month | Plus $20 /month | Pro 5x $100 /month | Pro 20x $200 /month | Business / Enterprise Not published | API Key Pay only for the tokens Codex uses, based on API pricing Strengths on file: Runs one agent across CLI, IDE, desktop, and cloud — work moves between surfaces.; Strong GitHub integration: a cloud code review on a pull request that catches real bugs and proposes fixes.; Reads the shared AGENTS.md standard and supports MCP tools.; Included with ChatGPT plans, from the free tier up through Pro. Limits on file: Locked to OpenAI's GPT models — no Claude or Gemini out of the box.; Fewer built-in customization primitives (hooks, sub-agents) than some rivals. Codex is OpenAI's agentic coding tool. It runs the same agent across a terminal CLI, an IDE extension, a desktop app, and a cloud environment, so work started in one surface can continue in another. It reads the shared AGENTS.md instructions file, speaks the Model Context Protocol (MCP), and can spin up a cloud code review on a GitHub pull request that comments inline and proposes fixes. Codex runs on OpenAI's GPT models — currently GPT-5.5 as the recommended frontier model — and is included with ChatGPT plans, from the free tier up through Pro.
Change log
- 2026-08-21 — First published, compiled from the catalogue entries for Claude Code and Codex.