Claude Code
by Anthropic
Codex
by OpenAI
Claude Code vs Codex
Claude Code from Anthropic and Codex from OpenAI are the two coding agents sold with a chat subscription. Here is how they compare on price, limits, models and features, checked against both vendors' own pages on November 20, 2026.
Prices and plans checked
The short version
These two are close in shape: both are agents that run in a terminal, an IDE extension, a desktop app and the cloud, and both come with a $20 subscription. The practical differences are the model family, how limits are published, the free tier, and the top price.
Pick Claude Code if
- You prefer Claude models for coding, or already pay for Claude Pro, Max, Ultra or a Team seat
- You want hooks, skills, subagents and scheduled tasks configured in your repo
- You want to bill through Amazon Bedrock, Google Cloud or Microsoft Foundry
- You want managed pull request review on a Team plan
Pick Codex if
- You prefer OpenAI models, or already pay for ChatGPT Plus, Pro or Business
- You want to try a coding agent on a free or $8 plan first
- You want an open-source CLI: OpenAI publishes the Codex CLI on GitHub
- You want published per-model usage estimates before you buy
They can share a repository: Claude Code can read an AGENTS.md file in place of CLAUDE.md, and both have VS Code extensions, so some teams keep both and use whichever suits the task.
Claude Code and Codex, side by side
- What it is
-
Claude CodeAn agentic coding tool that reads your codebase, edits files and runs commandsCodexOpenAI's coding agent, part of ChatGPT, with a terminal CLI, IDE extension and cloud tasks
- Where it runs
-
Claude CodeTerminal, VS Code extension, JetBrains plugin, desktop app, web and the Claude mobile appCodexTerminal CLI, IDE extension, ChatGPT desktop app, web (Codex Cloud) and the ChatGPT mobile app
- Starting price
-
Claude CodeClaude Pro at $20/month, or $17/month billed annuallyCodexChatGPT Plus at $20/month for the CLI, IDE extension, web and iOS. Limited access on Free ($0) and Go ($8)
- Top individual plan
-
Claude CodeUltra at $200/monthCodexPro at $100, $200 or $500/month
- Free tier
-
Claude CodeNo. Claude Code is not on the Claude Free planCodexYes, limited. Free and Go get GPT-6 Luna in the desktop app, subject to rollout
- Models
-
Claude CodeClaude only: Opus 5.5 by default, plus Sonnet 5.5, Haiku 5.5 and FableCodexOpenAI GPT-6 family: GPT-6.2 Astra and Sol, GPT-6.1 Sol and Terra, GPT-6 Astra and Luna. Availability varies by plan
- How usage is metered
-
Claude CodePlan limits per 5 hours and per week, shared with Claude chat. Max is 5x Pro; Ultra is 10x Pro per weekCodexMessages per 5 hours that vary by model, and weekly limits may also apply. Pro plans currently have no 5-hour limit
- Agent features
-
Claude CodeSubagents, background agents, cloud sessions, scheduled tasks, hooks, skills and MCPCodexSubagents, cloud tasks, scheduled tasks, skills and plugins, MCP, sandboxing and approval settings
- Code review
-
Claude Code/code-review in a session on any plan; managed GitHub Code Review on Team and EnterpriseCodexReview from the CLI; automatic GitHub code review as a cloud feature from Plus up
- CI and integrations
-
Claude Codeclaude -p, GitHub Actions, GitLab CI/CD, Slack, ChromeCodexcodex exec, GitHub Action, SDK, Slack, Linear, GitHub; GitLab in beta
- Team plans
-
Claude CodeTeam seats at $25 Standard or $125 Premium per month ($20 or $100 billed annually)CodexChatGPT Business at $25 per user per month, or $20 billed annually, 2 or more users
- Best for
-
Claude CodeDevelopers and teams who want Claude models and deep per-repo customizationCodexDevelopers and teams who want OpenAI models, a free way to start, or an open-source CLI
Claude Code vs Codex at a glance
Unlike Cursor and Claude Code, these two are the same kind of product. Each is an agent that reads your repository, edits files and runs commands, and each is sold as part of a chat subscription rather than on its own.
Claude Code runs in the terminal, in VS Code and JetBrains IDEs, in the Claude desktop app, in the browser at claude.ai/code and in the Claude mobile app. Codex runs as a CLI, as an IDE extension for VS Code, Cursor and Windsurf (with separate integrations for JetBrains IDEs and Xcode), in the ChatGPT desktop app for macOS, Windows and Linux, on the web as Codex Cloud, and in the ChatGPT mobile app.
OpenAI's docs now present Codex as part of ChatGPT, and the same documentation covers general work as well as coding. The pieces developers usually mean by "Codex" are the CLI, the IDE extension and Codex Cloud.
Because the feature lists are so similar, most people choose on three things: which models they prefer, what the subscription they already pay for includes, and how the limits fit their workload.
Codex pricing versus Claude Code pricing
Neither product has its own price. Codex is included in ChatGPT plans and Claude Code is included in Claude plans. Both can also be billed per token with an API key.
Codex has the wider range. OpenAI lists limited Codex access on ChatGPT Free and on Go at $8 a month, full access from Plus at $20, and three Pro tiers at $100, $200 and $500 a month. Claude Code starts at Pro for $20 and tops out at Ultra for $200; it is not available on the Claude Free plan.
Prices below are monthly list prices in USD before tax, from claude.com/pricing and OpenAI's Codex pricing page.
| Price point | Claude Code | Codex |
|---|---|---|
| $0 | Not available | ChatGPT Free: quick coding tasks with GPT-6 Luna in the desktop app, subject to rollout |
| $8 | No plan at this price | ChatGPT Go: lightweight coding tasks with GPT-6 Luna in the desktop app, subject to rollout |
| $20 | Claude Pro ($17 billed annually): Claude Code with Pro limits per 5 hours and per week | ChatGPT Plus: Codex on the web, CLI, IDE extension and iOS, plus automatic code review and Slack |
| $100 | Max: 5x Pro usage, $100 a month in API credits | Pro $100: everything in Plus, currently no 5-hour limit |
| $200 | Ultra: 10x Pro usage per week, no 5-hour limit for verified accounts, $200 a month in API credits | Pro $200: a higher Pro usage tier |
| $500 | No plan at this price | Pro $500: the highest Pro tier, with Ultrafast mode |
| Team seat | $25 Standard or $125 Premium ($20 or $100 billed annually) | ChatGPT Business: $25 per user, or $20 billed annually, 2 or more users |
| API key | Per token: Opus 5.5 $4 input / $20 output, Sonnet 5.5 $2 / $10 per million tokens | Per token: GPT-6.2 Astra $8 input / $40 output, GPT-6.2 Sol $2 / $10 per million tokens |
Usage limits: how each one counts
Claude Code uses your Claude plan limits. They reset on a 5-hour window and a weekly window and are shared with Claude chat. Anthropic states them as multiples: Max is 5x Pro, and Ultra is 10x Pro per week with no 5-hour limit for verified accounts. It does not publish message or token counts, and says how fast you reach a limit depends on the model and the length of the session.
Codex publishes estimates. For Plus and standard Business seats, OpenAI lists a range of local messages per 5 hours for each model: for example 17 to 180 on GPT-6.2 Sol, 6 to 50 on GPT-6.2 Astra and 350 to 3,000 on GPT-6 Luna. OpenAI calls these estimates rather than fixed limits, says weekly limits may also apply, and notes that cloud tasks may use more of the allowance than local messages. Pro plans currently have no 5-hour limit. Faster speed settings use the allowance at 2x or more.
Past the limit, both let you pay to keep going. Claude subscribers can turn on usage credits and continue at API rates up to a monthly spend limit. ChatGPT Plus and Pro users can buy ChatGPT credits, which Codex spends at a per-model rate.
For per-token budgeting, Anthropic gives one real-world figure: across enterprise deployments, API-billed Claude Code averages about $13 per developer per active day and $150 to $250 per developer per month. OpenAI's Codex pricing page does not give an equivalent average, only that a typical GPT-5.6 Sol task may use 2 to 15 credits.
Models
Claude Code runs Anthropic's models. On subscription plans it defaults to Opus 5.5, and you can switch to Sonnet 5.5, Haiku 5.5 or Fable with /model. Fable is included on Max and Ultra for up to half of your weekly limit; on Pro it is paid for with usage credits.
Codex runs OpenAI's models. The Codex docs list GPT-6.2 Astra for demanding multi-step work, GPT-6.2 Sol for everyday coding, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Astra and GPT-6 Luna as the most efficient option. Which of these you get depends on the plan: Free and Go are listed with GPT-6 Luna only, and the Plus plan summary names GPT-6.1 Sol and GPT-6 Luna, so check the model picker on your own plan.
Both can point at other endpoints. Claude Code supports Amazon Bedrock, Google Cloud's Agent Platform, Microsoft Foundry and LLM gateways. Codex supports custom model providers that expose a Responses API-compatible endpoint.
We do not quote benchmark scores here. Both vendors release new models often, and the better test is a week of your own tasks on each.
Agent features compared
The two feature lists read alike. These are the differences that are documented.
- Cloud work. Claude Code cloud sessions run on Anthropic-managed VMs on Pro, Max, Ultra and Team plans, count toward your plan limits, and can be pulled into your terminal with claude --teleport. Codex Cloud runs each task in its own workspace, keeps going while your computer is asleep, and can be started from the web, desktop app, mobile app, CLI or Slack. Both need you to sign in with the subscription account, not an API key.
- Scripting and CI. Claude Code has claude -p plus GitHub Actions and GitLab CI/CD integrations. Codex has codex exec, a GitHub Action and an SDK.
- Pull request review. Claude Code can auto-fix pull requests from a cloud session and offers a managed GitHub Code Review service on Team and Enterprise. Codex lists automatic GitHub code review as a cloud feature on Plus and above.
- Customization. Claude Code uses CLAUDE.md, skills, hooks, plugins and MCP servers, and can read an AGENTS.md file instead of CLAUDE.md. Codex uses AGENTS.md, skills, plugins and MCP servers, with sandboxing and permission profiles to control what it can run.
- Parallel work. Both support subagents. Claude Code also has background agents and experimental agent teams; the Codex desktop app is built around running several projects at once.
- Open source. OpenAI publishes the Codex CLI, SDK and app server on GitHub; the IDE extension and Codex Cloud are not open source.
- Scheduling. Both have scheduled tasks. Claude Code can run them in the cloud or on your desktop; Codex schedules them from the ChatGPT desktop app.
Which should you pick?
Pick Claude Code if you prefer Claude models, already have a Claude subscription, or want to run the agent through AWS, Google Cloud or Azure accounts. See Claude Code pricing for the plan-by-plan detail.
Pick Codex if you prefer OpenAI models, already pay for ChatGPT, want to start free, or care that the CLI is open source. If you need more than a $200 plan provides, Codex is also the only one of the two with a higher subscription tier.
If you have neither subscription, both cost $20 a month to try properly, and both install in a few minutes. Run the same two or three real tasks from your backlog through each before committing to a $100 or $200 plan.
For a team, compare Claude Team seats ($25 Standard, $125 Premium) with ChatGPT Business ($25 per user) on the same basis: who needs heavy agent usage and who mostly needs chat.
Pricing in detail
Claude Code plans
| Plan | Price | What you get |
|---|---|---|
| Free | $0 | Claude chat only. Claude Code is not included. |
| Pro | $20/month, or $17/month billed annually ($200) | Claude Code with Pro usage limits per 5 hours and per week. Opus, Sonnet and Haiku. Fable paid with usage credits. |
| Max | $100/month | 5x Pro usage per 5 hours and per week. Fable included for up to half your weekly limit. $100 a month in API credits. Priority access at high-traffic times. |
| Ultra | $200/month | 10x Pro usage per week with no 5-hour limit for verified accounts. $200 a month in API credits. |
| Team, Standard seat | $25/seat/month, or $20 billed annually | Claude Code with more usage than Pro. Central billing, SSO, admin controls. Teams of 2 to 150. |
| Team, Premium seat | $125/seat/month, or $100 billed annually | 5x the usage of a Standard seat. Fable for up to half of weekly limits. |
| Enterprise | $20/seat/month billed annually, plus usage at API rates | 20-seat minimum. Spend limits, SCIM, audit logs, Compliance API. |
| API (Claude Console) | Per token | No subscription. Opus 5.5 is $4 input / $20 output per million tokens; Sonnet 5.5 is $2 / $10. |
Source: claude.com/pricing and Anthropic's API pricing docs, checked November 20, 2026. Plan limits are shared between Claude Code and Claude chat. The monthly API credits on Max, Ultra and Team are for the Claude Platform (API), which is separate from your plan limits. Anthropic's $200 plan was previously sold as Max 20x.
Codex plans
| Plan | Price | What you get |
|---|---|---|
| Free | $0 | Quick coding tasks. GPT-6 Luna at Standard speed in the desktop app, subject to rollout. |
| Go | $8/month | Lightweight coding tasks. GPT-6 Luna at Standard speed in the desktop app, subject to rollout. |
| Plus | $20/month | Codex on the web, in the CLI, in the IDE extension and on iOS. Cloud integrations such as automatic code review and Slack. Extend usage with ChatGPT credits. |
| Pro | $100, $200 or $500/month | Everything in Plus at three usage tiers. Currently no 5-hour limit. Ultrafast mode on the $500 tier. |
| Business | $25/user/month, or $20 billed annually | 2 or more users. Codex across desktop and mobile apps, larger cloud VMs, SAML SSO and MFA, no training on business data by default. Extend usage with credits. |
| Enterprise and Edu | Contact sales | Everything in Business plus priority processing, SCIM, EKM, role-based access, Compliance API, data retention and residency controls. |
| API key | Per token at API rates | Codex in the CLI, SDK or IDE extension. No cloud features such as GitHub code review or Slack. |
Source: OpenAI's Codex pricing page, checked November 20, 2026. Usage on Plus is published as estimated messages per 5 hours by model, and weekly limits may also apply. OpenAI does not state on that page how much more usage each Pro tier includes. For usage-based billing it lists GPT-6.2 Astra at $8 input and $40 output per million tokens and GPT-6.2 Sol at $2 and $10.
Ways to pay less for either one
Claude Code pricing
Subscription versus API billing, and which is cheaper for you.
Claude pricing
All Claude plans, including the ones that include Claude Code.
OpenAI API pricing
Per-token rates if you run Codex on an API key.
OpenAI for Startups
How OpenAI startup credits work and who can get them.
Claude for Startups
Closed to new applicants since October 13, 2026. We track when it reopens.
How to install Claude Code
Setup on Mac, Windows and Linux.
Questions people ask
Is Codex better than Claude Code?
How much does Codex cost?
Is Codex free?
Is Claude Code free?
Is Codex included with ChatGPT Plus?
What are the usage limits on Codex and Claude Code?
Can I use Claude Code and Codex in the same project?
AGENTS.md file in place of CLAUDE.md, so one instructions file can serve both, and both have VS Code extensions that can be installed side by side.