๐ฏ Quick Answer: Codex Pricing in 2026
Codex does not have its own subscription. It is included in ChatGPT Plus ($20/month), Pro ($100, $200 or $500/month), Business and Enterprise, with a limited amount on Free. You can also run it with an API key and pay per token at API prices. Extra usage can be bought with ChatGPT credits.
Best for: Developers deciding between a flat ChatGPT plan and pay-as-you-go API billing.
Three ways to pay for Codex
1. A ChatGPT plan. Plus, Pro, Business, Enterprise and Edu all include Codex. Plus costs $20/month, and Pro tiers run $100, $200 and $500/month with more usage as you go up. See the full breakdown in our ChatGPT pricing guide.
2. ChatGPT credits. When you hit your plan limit, credits extend usage on Plus, Pro, Business and Enterprise plans.
3. An API key. Use Codex in the CLI, SDK or IDE extension and pay only for the tokens it uses, at OpenAI API prices. Cloud features like GitHub code review and Slack are not part of API-key use, and the models available follow what your key can access.
What changes by plan
| Option | Cost | Codex access |
|---|---|---|
| Free | $0 | Limited Codex access |
| Plus | $20/month | Codex on web, CLI, IDE and iOS; cloud integrations; credits to extend |
| Pro | $100, $200 or $500/month | More included usage than Plus; model availability and usage depend on plan and workspace |
| Business | $25/user/month ($20 annual) Standard; $125 ($100 annual) Premium | Codex in both seat types; Premium has 5x the usage of Standard |
| Enterprise | Custom | Priority processing, SCIM, EKM, RBAC, audit logs |
| API key | Pay per token | CLI, SDK, IDE; no cloud-based features |
ChatGPT Work shares your Codex usage
OpenAI says ChatGPT Work and Codex share usage, so Work tasks inside ChatGPT draw on the same pricing, credits and limits. If you use both, your Codex allowance runs down faster.
Which should you pick?
Start with Plus if you code a few hours a week. Move to Pro when you hit limits daily. Use an API key if you want predictable per-token billing or run Codex in CI. Compare with Cursor pricing if you prefer an editor-first tool.
What people mean when they search Codex pricing
There is no standalone Codex price tag, which is why the search is confusing. Codex is bundled into ChatGPT plans, can be topped up with ChatGPT credits, or can be run on an API key where you pay per token. So the real question is which of those three billing routes fits how you code, not what Codex costs on its own.
Plan route vs API-key route
A ChatGPT plan gives you a flat monthly price and the full Codex experience, including cloud features like GitHub code review and Slack. Plus is $20 per month. The Pro tiers are $100, $200 and $500 per month and add more usage as you go up. The tradeoff is that usage is capped by plan, and ChatGPT Work shares that same allowance.
An API key means you pay only for tokens used, at OpenAI API prices. It works in the CLI, SDK and IDE extension, which makes it a good fit for CI jobs or occasional use where a monthly plan would sit idle. It does not include cloud features such as GitHub code review or Slack, and the models you can use follow what your key can access. Because the bill scales with usage, a heavy month can cost more than a flat plan, so set a spending limit.
How to decide: three typical users
A weekend builder who codes a few hours a week should start on Plus at $20 and see whether the limit ever bites. A full-time developer who hits limits most days should look at the Pro tiers, picking the lowest one that stops the interruptions rather than jumping straight to $500. A team automating code review or running agents in pipelines should compare Business or Enterprise with an API key, since the right answer depends on whether they need the cloud features.
Ways to avoid overpaying
Buy credits instead of upgrading a tier when you only occasionally run out. Keep tasks scoped so each run uses less of your allowance. Remember that anything you do in ChatGPT Work counts against the same pool. And if you already pay for another coding tool, compare Cursor pricing and Devin pricing before adding Codex on top.
Codex billing routes in plain terms
Codex has no standalone subscription. The plan route bundles Codex into ChatGPT Plus at $20 per month, Pro at $100, $200 or $500 per month, Business Standard at $25 monthly or $20 annually per seat, Premium at $125 monthly or $100 annually, and custom Enterprise. The plan route includes cloud features such as GitHub code review and Slack, plus web, CLI, IDE and iOS surfaces on paid plans. The API-key route bills per token at OpenAI API prices for CLI, SDK and IDE use, with no cloud features, and the available models follow the key. Credits extend plan usage when limits hit on Plus, Pro, Business and Enterprise.
Annual math that matters
Plus at $20 is $240 over twelve months. Pro 100 at $100 is $1,200 per year, Pro 200 is $2,400, and Pro 500 is $6,000. The jump from Plus to Pro 100 is $80 per month or $960 per year, so only make it when Plus caps interrupt paid work most days. The jump from Pro 100 to Pro 200 is another $100 per month, and Pro 200 to Pro 500 is $300 more. For occasional overflow, credits beat a permanent tier jump. For CI jobs or occasional use where a monthly plan would sit idle, an API key with a spending limit is usually cheaper.
Teams and shared usage
Business needs at least two seats and can mix Standard and Premium, so reserve Premium for high-volume developers since it carries five times Standard usage. ChatGPT Work shares the same Codex allowance, so teams using both should expect faster burn. Enterprise adds priority processing, SCIM, enterprise key management, role controls and audit logs for larger deployments. Compare Cursor pricing and Devin pricing before stacking another coding subscription on top of Codex.
Which Codex models are available: GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Sol and GPT-6 Luna
OpenAI's current Work and Codex guide lists GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Sol and GPT-6 Luna. Which models appear depends on your account, plan, workspace permissions and rollout. OpenAI says the Work and Codex allowance is not a fixed message count: it varies with your plan, model, task and settings. Check Settings โ Usage for the allowance and reset times shown for your account.
What actually burns your Codex allowance
OpenAI states that Codex usage depends on the model, the size and complexity of the task, and whether work runs locally or in the cloud. Small scripts or routine functions may use only a fraction of your allowance, while larger codebases, long-running tasks, and sessions that hold more context use more. Prompt length alone is not a reliable estimate because context, reasoning, tool use, retrieval and caching all affect usage. Depending on your plan, limits may apply over five-hour and weekly windows; Pro 100, Pro 200 and Pro 500 currently have no five-hour limit, but still have an included usage allowance. Scope tasks tightly and use a model suited to the work. Codex usage is also shared with other agentic features once their pricing is effective, currently including ChatGPT for Excel on Plus and Pro, so heavy spreadsheet agent use draws from the same pool.
Enterprise, education and invite notes
Enterprise and education buyers have two routes: flexible pricing with no fixed rate limits where usage scales with credits, or standard per-seat plans that carry roughly Plus-level limits for most features. The Enterprise route adds priority processing, SCIM, enterprise key management, role controls, domain verification, audit logs and compliance API monitoring, plus data retention and residency controls. Invites work differently by workspace: eligible personal plans and Business workspaces have separate referral rewards and eligibility rules shown in the invite dialog, while Enterprise has no referrals. Larger virtual machines for faster cloud chats sit on the Business and Enterprise side rather than the individual plans.
Credits, limits and checking the current page
ChatGPT credits can extend eligible Codex usage after a plan allowance is used. OpenAI's current usage guidance says allowances vary by plan, seat, model, task and settings; some plans have five-hour and weekly limits, while Pro 100, Pro 200 and Pro 500 have no five-hour limit but still have an overall allowance. Check Settings โ Usage for the current allowance and reset times on your account, since availability and limits can differ by workspace and rollout.
Already considering Claude Code? Our Codex vs Claude Code comparison explains which to try first and how to compare them on your own app.
Frequently Asked Questions
Is Codex free?
Free ChatGPT accounts include limited Codex access. For regular use you need a paid ChatGPT plan or an API key.
How much does Codex cost?
There is no separate Codex subscription. It is included in ChatGPT Plus at $20/month, Pro at $100, $200 or $500/month, and Business and Enterprise plans. With an API key you pay per token.
Can I buy more Codex usage?
Yes. ChatGPT credits extend usage once you reach your plan limit on Plus, Pro, Business and Enterprise.
Does Codex with an API key include cloud features?
No. API-key use covers the CLI, SDK and IDE extension, but not cloud-based features like GitHub code review or Slack integration.
Should I use a ChatGPT plan or an API key for Codex?
Use a plan for a flat price and cloud features like GitHub code review. Use an API key for pay-per-token billing in the CLI, SDK, IDE extension or CI.
Does ChatGPT Work use my Codex allowance?
Yes. OpenAI says ChatGPT Work and Codex share usage, so heavy use of one reduces what is left for the other.
Do I need a Pro plan for Codex?
Not necessarily. Plus at $20 per month includes Codex. Move to a Pro tier only if you hit limits regularly.
Which models can I use with an API key?
The models available follow what your API key can access.
Advertiser disclosure: some links on this website are affiliate links, meaning No Code MBA may make a commission if you click through and purchase.