Guide · Codex
Codex usage limits, explained
Your ChatGPT plan sets the allowance, and a five-hour and a weekly window can both apply. Here’s how they work and where to check them.
By Thomas Kanze
Last checked October 9, 2026. Limits change; the sources linked below are the final word.
The short version
Codex is included with ChatGPT plans, and your plan sets how much you can use. Depending on the plan, limits can apply over a five-hour window and a weekly window, and you need room in both to keep going. OpenAI says Pro plans currently have no five-hour limit. Check yours in the Codex usage dashboard or with /status in the Codex CLI.
Which plans include Codex
OpenAI’s Codex pricing page lists Codex as included with ChatGPT Free, Go, Plus, Pro, Business, Edu and Enterprise plans. Enterprise and Edu workspaces with flexible pricing have no fixed rate limits; usage draws on credits instead.
Local messages (CLI, IDE extension, app) and cloud chats share the same plan allowance. OpenAI notes that cloud tasks may use more of it than local messages.
Plans that include Codex, as listed on OpenAI’s Codex pricing page (October 9, 2026)
| Plan | Listed price | Five-hour window |
|---|---|---|
| Free | $0 | Plan allowance applies |
| Go | $8/month | Plan allowance applies |
| Plus | $20/month | Estimates published per five hours |
| Pro | From $100/month | None currently, per OpenAI |
| Business | $25/user monthly, $20 billed yearly | Estimates published for Standard Business |
| Enterprise / Edu | Contact sales | Flexible pricing: no fixed rate limits |
Prices and windows change; the Codex pricing page is the final word.
How the windows work
The five-hour window
OpenAI’s help center says a new five-hour window starts when you send your first message after the previous window ends. So the window follows your activity, not a fixed time of day. You can also hit the five-hour limit before five hours have passed, even with weekly usage left.
The weekly window
OpenAI says weekly limits may also apply. The help center says the weekly period follows your previous weekly reset date, and using a banked or purchased reset can move that date. Your dashboard shows the current reset times.
Not a message count
OpenAI publishes estimated message ranges for some plans and models, but says they aren’t fixed limits. Usage depends on the model, where the task runs, how complex it is, context, reasoning, speed and tools. A long-running task can use far more than a short request. The ranges also change with each model release, so we don’t repeat them here.
How to check your Codex usage
- Usage dashboard: chatgpt.com/codex/settings/usage shows your current limits and reset times.
/statusin the CLI: OpenAI’s pricing page says/statusshows your remaining limits during an active Codex CLI session. The slash-command reference describes it as session configuration and token usage. If the numbers look off, trust the dashboard./statusline: adds a rate-limits item to the Codex TUI status line so the number stays on screen./usage: shows daily, weekly or cumulative token activity, and lets you redeem an earned reset if you have one.
If you also use Claude Code, Quotchi shows both tools’ usage and reset times side by side in the Mac menu bar. It uses each tool’s existing sign-in, so there’s nothing new to log in to.
What to do when you hit a limit
- Let the current turn finish. OpenAI says an agent that hits your limit mid-turn can keep working on that turn, subject to fair-use limits.
- Add credits. Plus and Pro users who reach their limit can buy credits to keep working. Your plan’s included usage is used first.
- Use a reset, if you have one. Some accounts have banked or purchasable resets. A banked reset refreshes the five-hour and weekly windows and changes your weekly reset date.
- Use an API key for local work. OpenAI says anyone can run extra local chats with an API key, billed at standard API rates.
- Wait for the reset time on your dashboard, or upgrade your plan.
Don’t expect support to clear it for you. OpenAI’s help center says support doesn’t reset ChatGPT or Codex usage limits.
Quotchi is a small Mac menu bar app for Codex and Claude Code limits: current usage, reset times and gentle alerts. It’s Free, local-first, and needs no Quotchi account.
See how Quotchi worksWhat uses more of your allowance
OpenAI is explicit that prompt length alone isn’t a reliable estimate. The things that move the needle:
- Model choice, context and reasoning. Bigger models, longer context and more reasoning cost more per task.
- Tool use, retrieval and caching. All listed by OpenAI as affecting usage.
- Where it runs. Cloud tasks may use more of your allowance than local messages.
- Speed modes. Fast and Ultrafast use included limits and paid credits at different rates.
- Images. OpenAI says image generations use included limits about three to five times faster on average.
A worked example
Plus, one busy afternoon
You start a cloud task at 1pm and keep chatting with the CLI locally. By 3pm the dashboard shows the current five-hour window nearly used, with a reset time that evening. If the window runs out mid-turn, the cloud task can still finish that turn: OpenAI lets an agent complete a turn it started, subject to fair use. You switch the remaining work to small local tasks. Then you decide whether the rest is worth buying credits for or can wait for the reset.
Had you been on Pro, there’d be no five-hour window to hit right now, per OpenAI, though your included allowance and any weekly limit still apply.
Plan around your Codex windows
- Check the dashboard before cloud-heavy work. Cloud tasks may use more allowance than local messages, so start them with room to spare.
- Go local and small when the window is tight. Local and cloud share one allowance, but cloud tasks may use more of it, and a small local task is easier to fit before the reset.
- Save banked resets for real blocks. A reset refreshes both windows but moves your weekly reset date, so it’s a tool for a deadline, not a routine.
- Keep the number on screen.
/statuslinewith the rate-limits item puts it in the Codex status line while you work. - Have a second option. Credits, an API key for local chats, or another tool with its own allowance, decided before you need it.
For each way to read your numbers, see how to check Codex usage.
Common mistakes and myths
“The published message counts are my quota.”
OpenAI says the estimates aren’t fixed message limits and points you to the dashboard for current limits.
“Pro has unlimited Codex.”
OpenAI says Pro currently has no five-hour limit. That’s not the same as no allowance.
“Support can reset my limit.”
OpenAI’s help center says support doesn’t reset ChatGPT or Codex usage limits.
“Local use doesn’t touch my cloud allowance.”
Local messages and cloud chats share the same plan allowance.
What changed recently
As of October 9, 2026: OpenAI’s Codex pricing page says GPT-5.5 retires from ChatGPT, ChatGPT Work and Codex on all plans on October 14, 2026 (the API isn’t affected), and that GPT-5.6 Sol’s promotional pricing runs at least through November 21, 2026. It also says Codex usage limits are shared with other agentic features once their pricing takes effect.
Sources
Official pages we read on October 9, 2026. If anything here disagrees with them, they win.
FAQ
Does Codex have a five-hour limit?
Depending on your plan, yes. Limits may apply over a five-hour window and a weekly window. OpenAI says Pro plans currently have no five-hour limit, though they still have an included allowance.
When does the Codex five-hour window start?
According to OpenAI, a new five-hour window starts when you send your first message after the previous window ends.
How do I check my Codex usage?
Open the usage dashboard at chatgpt.com/codex/settings/usage, or run /status in an active Codex CLI session.
Do the CLI and Codex cloud share limits?
Yes. OpenAI says local messages and cloud chats share your plan allowance, and cloud tasks may use more of it.
Can OpenAI support reset my Codex limit?
No. OpenAI says support doesn’t reset ChatGPT or Codex usage limits. Wait for the reset, add credits, or use a reset option shown in your account.
Which ChatGPT plans include Codex?
OpenAI lists Free, Go, Plus, Pro, Business, Edu and Enterprise.
Do Fast or Ultrafast modes use more of my limit?
OpenAI says they use included subscription limits and paid credits at different rates. Check the pricing page for current rates.
Can an agent finish its work if I hit the limit mid-task?
OpenAI says that if you reach your limit during an active turn, the agent can keep working on that turn, subject to fair-use limits.
Where can I learn every way to check usage?
See our guide to checking Codex usage: the dashboard, /status, /statusline and /usage.