Codex is included in ChatGPT plans, so its price is the price of your ChatGPT plan: as of October 2026, Free ($0) and Go ($8 a month) get limited desktop access, Plus is $20 a month, Pro starts at $100 a month, and Business is $20 per user per month billed annually. You can also skip the subscription and pay per token with an OpenAI API key.
All figures below come from OpenAI's Codex pricing page, read on 24 October 2026. Prices and limits change often, so check that page before you buy.
Which ChatGPT plans include Codex?
| Plan | Price (as of October 2026) | What it includes for Codex |
|---|---|---|
| Free | $0 a month | GPT-6 Luna at Standard speed in the desktop app, subject to rollout |
| Go | $8 a month | GPT-6 Luna at Standard speed in the desktop app, subject to rollout |
| Plus | $20 a month | Codex on the web, CLI, IDE extension and iOS; GPT-6.1 Sol and GPT-6 Luna; code review and Slack integrations; extra usage with credits |
| Pro | From $100 a month ($100, $200 or $500 plans) | Everything in Plus; no five-hour limit currently; Ultrafast mode on Pro $500 |
| Business | $20 per user a month billed annually (2+ users), $25 billed monthly | Desktop and mobile apps, larger cloud machines, admin controls, SAML SSO, no training on business data by default |
| Enterprise and Edu | Contact sales | Everything in Business plus priority processing, SCIM, EKM, RBAC, audit logs and data residency controls |
| API key | API token pricing | CLI, SDK, IDE extension and desktop local chats; no cloud features |
One detail is easy to miss: OpenAI says "ChatGPT Work and Codex share usage." If you also use ChatGPT Work for research or documents, it draws from the same allowance as your coding.
For other ChatGPT features in each plan, see chatgpt.com/pricing.
How many Codex messages do you get?
OpenAI publishes estimates rather than fixed numbers. These are local messages per five-hour window on Plus and Standard Business, as of October 2026:
| Model | Plus | Standard Business |
|---|---|---|
| GPT-6 Astra | 5 to 45 | 5 to 45 |
| GPT-6.1 Sol | 15 to 160 | 15 to 160 |
| GPT-6 Sol | 15 to 150 | 15 to 150 |
| GPT-6 Luna | 350 to 3,000 | 350 to 3,000 |
The ranges are wide because usage depends on the work. OpenAI's page explains that "Model choice, context, reasoning, tool use, retrieval, and caching all affect usage, so prompt length alone isn't a reliable estimate." Other rules:
- Pro plans currently have no five-hour limit.
- Weekly limits may also apply on top of the five-hour window.
- Local and cloud share one allowance, and cloud tasks may use more of it than local messages.
- Enterprise and Edu without flexible pricing get the same per-seat limits as Plus for most features. With flexible pricing, there are no fixed rate limits and usage scales with credits.
- Image generation uses included limits about 3 to 5 times faster on average, and it is not available on the Free plan.
How do Codex credits work?
Credits pay for usage beyond your included limits on credit-based plans. Plus and Pro users can buy credits when they hit a limit instead of upgrading, and Business, Enterprise and Edu workspaces on flexible pricing can buy workspace credits.
Credits are charged per million tokens at Standard speed. A selection of OpenAI's rates as of October 2026:
| Model | Input | Cached input | Output |
|---|---|---|---|
| GPT-6.1 Sol | 50 credits | 2.5 credits | 250 credits |
| GPT-6 Sol | 50 credits | 5 credits | 250 credits |
| GPT-6 Luna | 2.5 credits | 0.25 credits | 12.5 credits |
| GPT-6 Astra | 250 credits | 25 credits | 1,250 credits |
Codex credit billing has no separate cache-write charge. Credit purchase prices depend on your plan or agreement, and OpenAI notes that credit rates alone don't determine how much is included in your subscription. The full rate card, including older and specialist models, is on the pricing page.
How do Fast and Ultrafast modes affect usage?
Speed modes cost more than Standard for the same model:
| Speed mode | Included subscription usage | Purchased credits and Enterprise pay-as-you-go |
|---|---|---|
| Fast | 2.5x | 2x |
| GPT-6 Astra Ultrafast | 8x | 6x |
| GPT-6.1 Sol Ultrafast | 8x | 6x |
OpenAI's feature table lists Fast mode on Plus, Pro, Business, Enterprise and Edu, and API keys. Ultrafast is limited to Pro $500, eligible Enterprise and Edu plans, and API keys. OpenAI's speed docs list which models support each mode.
Voice in the desktop app uses your Codex budget at $0.05 per minute, or 1.25 credits per minute on credit-based workspaces. The model doing the task is billed separately.
Should you use a ChatGPT plan or an API key?
| ChatGPT plan | API key | |
|---|---|---|
| How you pay | Monthly subscription with included limits, plus optional credits | Per token at API pricing |
| Where Codex works | Desktop app, CLI, IDE extension, web, mobile, Codex cloud | CLI, SDK, IDE extension, desktop app local chats |
| Cloud features | GitHub code review, @codex delegation, Slack, Linear, cloud tasks |
Not available |
| Models | Models your plan includes | Models available to your key |
OpenAI describes the API key option as "Great for automation in shared environments like CI." That fits codex exec in a pipeline, where a personal subscription would be the wrong account to bill. For daily interactive work, a plan is usually simpler because usage is included. You can also mix them: OpenAI says all users can run extra local chats with an API key, charged at standard API rates.
The Codex CLI guide shows how to sign in with either method.
How do you make your Codex limits last longer?
- Check the dashboard. The usage dashboard shows current limits and reset times. In the CLI,
/statusshows what is left. - Pick a lighter model for light work. GPT-6 Luna's estimated message range is far higher than GPT-6.1 Sol's.
- Trim context. Be precise in prompts, share only relevant files, and keep AGENTS.md lean by nesting it in subfolders. See the AGENTS.md guide.
- Turn off unused MCP servers. OpenAI notes that every MCP server adds context and uses more of your limit. The Codex MCP guide shows how to disable one.
- Know what counts as Code Review. Code Review usage applies only to reviews run through GitHub, such as tagging
@codexon a pull request. Local reviews count toward your general limits.
One recent change: GPT-5.5 retired from Codex with ChatGPT sign-in on October 14, 2026, at noon PDT. The OpenAI API and Codex sessions signed in with an API key were not affected.
How does Codex pricing compare with Claude Code and Cursor?
Each tool has its own plans, limits and billing rules, so compare them on their own pages: the Claude Code pricing guide and the Cursor pricing guide. The Codex vs Claude Code comparison covers the rest of the differences.
Whichever tool you pay for, your prompts and skills don't need to be tied to it. Kipt keeps them as Markdown in a private GitHub repo you own and installs every skill into Codex or Claude Code with npx skills add https://github.com/<you>/<repo>.git --all -g.