# Claude Code vs Codex: compare the workflow, then try your task

> Claude Code and Codex both read a codebase, edit files and run commands, in the terminal, an editor or the browser. Both come with a $20 monthly plan: Claude Pro and ChatGPT Plus. Neither is better for every team, so run the same real task in both and compare the fix, your review time and the usage it took. Both can share notes and handoffs in 1folder.

*Picture: Claude Code and Codex side by side. Local coding work: Claude Code, Read a codebase, edit files and run commands; Codex, Inspect local files, edit code and run commands through Codex CLI. Where you work: Claude Code, Terminal, IDE, desktop and browser; Codex, CLI, IDE, web and app; by plan. Individual paid starting point: Claude Code, Pro: $20 monthly or $200 annually; Codex, Plus: $20 monthly.* Claude Code and Codex side by side: what they do, where they run and what the first paid plan costs, from each company’s own pages on November 11, 2026.

## What their own documentation establishes

| | Claude Code | Codex |
|---|---|---|
| Local coding work | Read a codebase, edit files and run commands | Inspect local files, edit code and run commands through Codex CLI |
| Where you work | Terminal, IDE, desktop app and browser; docs include VS Code and JetBrains | CLI, IDE extension, web and app experiences; availability depends on plan |
| Individual paid starting point | Claude Pro includes Claude Code: $20 monthly, or $200 billed annually | ChatGPT Plus includes Codex CLI, IDE and web: $20 monthly |
| Higher individual plans | Max $100 monthly; Ultra $200 monthly | Pro options at $100, $200 or $500 monthly |
| Check actual remaining usage | Run /usage to see plan usage bars and a usage breakdown | Use the usage dashboard or /status in Codex CLI |

Checked on November 11, 2026, from Claude Code docs, Claude’s pricing page and official OpenAI documentation. Prices are in USD; current plan pages and account usage displays take precedence.

## Which to pick

### Pick Claude Code if

- You already pay for Claude Pro, Max or Ultra: Claude Code comes with it.
- You need more than Pro’s usage: Max is $100 a month, Ultra $200 with no five-hour limit for verified accounts.
- On your own repository, its fixes are the ones you understand, can check and want to keep.

### Pick Codex if

- You already pay for ChatGPT Plus or Pro: Codex comes with it, on the web, CLI, IDE extension and iOS.
- You want to try it before paying: OpenAI lists light Codex use on Free and Go in the desktop app, subject to rollout.
- On the same test, its fixes are the ones you understand, can check and want to keep.

## Claude Code vs Codex pricing: compare access, not just a price

For one person paying monthly, Claude Pro and ChatGPT Plus are both $20, and each includes its coding tool. Claude Pro is also $200 a year paid up front. The same price doesn’t buy the same amount of coding: each plan has its own usage limits.

Above that, Claude lists Max at $100 and Ultra at $200 a month; OpenAI lists Pro at $100, $200 or $500 a month. OpenAI also lists Codex on Free and Go for light coding in the desktop app, subject to rollout, without the other places Plus gets.

Both companies also offer pay-as-you-go billing by usage, which works differently from a subscription. Compare the way you’ll actually sign in and pay. Prices change often; check the linked pages before you buy.

## Claude Code vs Codex usage limits are not fixed tasks per month

Claude’s pricing page describes a five-hour limit and a weekly limit. Max gives five times Pro’s usage. Ultra gives ten times Pro’s weekly usage and, for verified accounts, no five-hour limit. In Claude Code, /usage shows how much you have left.

OpenAI gives estimates of local messages per five hours for Plus, says Pro plans currently have no five-hour limit, and says weekly limits may also apply. It states that the estimates are not fixed message limits. The usage dashboard, or /status in Codex CLI, shows your own account.

So neither plan converts into “bugs fixed a month”. A one-line fix and a long debugging session are different jobs. In a trial, note for each tool whether the fix worked, how long you spent checking it and how much usage it took.

## Give both tools the same test

- In Claude Code: “Read CLAUDE.md first. Find why an empty search query produces no results. Explain the cause, add a regression test and make the smallest fix. Run the relevant tests. Report changed files, evidence and remaining risks.”
- In Codex: “Read AGENTS.md first. Find why an empty search query produces no results. Explain the cause, add a regression test and make the smallest fix. Run the relevant tests. Report changed files, evidence and remaining risks.”

## Judge the result on your repository

- Start each tool from its own copy of the same commit, with the same request.
- Check that the cause it names really explains the bug. A confident explanation isn’t evidence.
- Look for unrelated changes in the patch, and check that the new test fails without the fix.
- Run the checks yourself and note how long you spent reviewing or correcting each result.

## Use both on the same briefs and reports in 1folder

1. **Save the reviewed handover.** After connecting [Claude Code](https://1folderapp.com/integrations/claude-code-mcp.md) and [Codex](https://1folderapp.com/integrations/codex-mcp.md) to the same 1folder folder, ask Claude Code: “Save Search fix report in 1folder with the cause, changed files, tests run and open risks.” Keep application code in your normal repository.

   *Picture: In Claude Code, you type: “Save Search fix report in 1folder with the cause, changed files, tests run and open risks.” Claude Code saves Search fix report (a page saved by Claude Code, version 1) to 1folder.*
2. **Update it after a second review.** Ask Codex: “Read Search fix report in 1folder. Review the patch in this checkout, add your findings and update the same report.” The file shares the written handover; it does not transfer a coding session, repository or memory.

   *Picture: The versions of Search fix report: Added second-review findings (Codex, today); Recorded the fix and test evidence (Claude Code, earlier).*

## Questions

### Claude Code vs Codex: which is better?

It depends on your code and how you work, so this page doesn’t pick one. Both read code, edit files and run commands; they differ in plans, limits and where they run. Give both the same real task and compare the fix, the time you spent checking it and the usage it took.

### Is Codex cheaper than Claude Code?

Not in general. ChatGPT Plus and Claude Pro are both $20 a month, with different usage included. Codex on Free and Go, the higher plans and pay-as-you-go billing can each change the answer for you.

### Does either plan mean unlimited coding?

No. Both have usage limits. Claude Ultra and OpenAI’s Pro plans have no five-hour limit, but weekly limits still apply.

### Can I use Claude Code and Codex together?

Yes. Both can read an AGENTS.md file in the repository (Claude Code reads it in place of CLAUDE.md), so one file of project rules can serve both. For briefs and reports, both can [read and update the same files](https://1folderapp.com/integrations/claude-code-and-codex-together.md) in 1folder. Neither sees the other’s session.

## Sources

- [Claude Code docs: Overview](https://code.claude.com/docs/en/overview)
- [Claude: Pricing](https://claude.com/pricing)
- [Claude Code docs: How Claude remembers your project](https://code.claude.com/docs/en/memory)
- [OpenAI Codex docs: Custom instructions with AGENTS.md](https://developers.openai.com/codex/guides/agents-md)
- [Claude Code docs: Manage costs effectively](https://code.claude.com/docs/en/costs)
- [Official OpenAI documentation: Codex CLI](https://learn.chatgpt.com/docs/codex/cli)
- [Official OpenAI documentation: Pricing](https://learn.chatgpt.com/docs/pricing)

Checked on November 11, 2026.

## Read next

- [Use Claude Code and Codex together on the same files, with Cursor too](https://1folderapp.com/integrations/claude-code-and-codex-together.md): Use Claude Code, Codex and Cursor together with a saved handoff, code revision and checks. Connect each app once and keep the next agent’s starting point clear.
- [Claude Code MCP: add 1folder so your plans and handoffs are kept](https://1folderapp.com/integrations/claude-code-mcp.md): Add 1folder to Claude Code with one command. It saves plans, handoffs and reports you can open anywhere, and Codex, Cursor or ChatGPT can pick them up.
- [Codex MCP: add 1folder so Codex saves and reads your files](https://1folderapp.com/integrations/codex-mcp.md): Add 1folder to Codex with codex mcp add and sign in. Codex then saves handoffs and reports to your files, where Claude Code or ChatGPT can pick them up.
- [Claude Code handoff: pass work to a new session](https://1folderapp.com/how-to/claude-code-handoff.md): Write a Claude Code handoff with the revision, changed files, checks and next steps. Compare HANDOFF.md, resume and compact, and save it to 1folder to share.

[Add 1folder to your AI app, free](https://1folderapp.com/signin) · [All resources](https://1folderapp.com/resources.md)

Website: https://1folderapp.com/ · Contact: hello@1folderapp.com