Pricing
Codex "You've Hit Your Usage Limit. Try Again Later.": What It Means
Codex says 'You've hit your usage limit. Try again later.'? Which limit tripped, how to check /status, when it resets, and the legit ways to keep coding.
9 min read
Error wording checked Oct 30, 2026 in three issues on OpenAI’s official Codex GitHub repo (#57655, #36344, #49773) and re-read Oct 31, 2026. Limit, reset and credit rules checked Oct 31, 2026 in OpenAI’s Codex plan help article and credits help article; plan limits checked Oct 30, 2026 on Codex pricing. Every source is listed at the bottom. No affiliate links on this page (how we work).
The exact error: You've hit your usage limit. Try again later.
This is the message people are pasting into Google:
You've hit your usage limit. Try again later.
Source: openai/codex issue #57655 (opened Oct 28, 2026), verified Oct 30, 2026. It has a sibling that adds the paid way out, from issue #49773 (verified Oct 30, 2026):
You've hit your usage limit. Add credits to continue, or try again later.
The one-line answer (as of Oct 31, 2026): you’ve used up one of your plan’s Codex allowances, usually the five-hour window or the weekly limit, and Codex won’t start a new turn until it resets or you add credits. OpenAI’s help center says to “open Settings or your usage dashboard. Check which allowance is exhausted, your credit balance, and any reset time shown. In an active Codex CLI session, enter /status” (help article, verified Oct 31, 2026).
Full plan-by-plan tables live in Codex usage limits.
Which limit did you hit?
OpenAI publishes two windows: estimates “per five-hour period”, and “Weekly limits may also apply” (Codex pricing, verified Oct 30, 2026). Here’s who gets which, using only what OpenAI prints:
| Plan | Window | Resets | Your move |
|---|---|---|---|
| Free / Go | Not published | Not published | Wait or upgrade |
| Plus | 5-hour + weekly | Dashboard time | Credits or wait |
| Pro | Weekly may apply | Dashboard time | Credits or wait |
| Business (Std) | 5-hour + weekly | Dashboard time | Workspace credits |
| Business (Prem) | No 5-hour limit | Dashboard time | Workspace credits |
Sources: Codex pricing (Plus and Standard Business ranges, “Pro plans currently have no five-hour limit”, verified Oct 30, 2026); ChatGPT Business help (Premium seat “no five-hour usage limit”, verified Oct 30, 2026).
A few things the table can’t fit:
- No public reset hour. Your reset time is the one shown in your account. OpenAI publishes no weekly allowance size.
- The five-hour numbers are estimates, not caps. On Plus, OpenAI estimates 15–160 local messages per five hours on GPT-6.1 Sol, 35–285 on GPT-6.1 Terra, 5–45 on GPT-6 Astra and 350–3,000 on GPT-6 Luna. In its words, “These estimates are not fixed message limits” (verified Oct 30, 2026).
- Codex shares its pool. Codex, ChatGPT Work, ChatGPT for Excel and Workspace Agents “use a shared allowance and credit pool when those features are available on your plan” (help article, verified Oct 31, 2026). A long spreadsheet session in the afternoon can be why your evening coding session dies.
- Pro 200 changed on Oct 29, 2026. Grandfathered Pro 200 subscribers kept their old allowance through Oct 29, 2026, then moved to the lower allowance at the same $200 (ChatGPT Pro help, verified Oct 29, 2026). If you’re on that plan and started seeing this message on Oct 30, that’s the first thing I’d look at.
Check where you stand
Two places, as OpenAI documents them (verified Oct 31, 2026):
- In the CLI: type
/statusin an active Codex session. The refreshed CLI also lists “Usage stats: Review usage information in the CLI.” - In ChatGPT or the Codex app: Settings > Usage in ChatGPT, or Usage & Billing in the Codex app. That’s where your credit balance lives too (credits help).
Look for which allowance is used up (five-hour or weekly), the reset time and your credit balance. An empty five-hour bar means hours of waiting; an empty weekly one can mean days.
And one myth to kill early: a fresh chat won’t save you. “Starting a new chat does not reset your usage limits” (help article, verified Oct 31, 2026). Closing the tab and opening a new one is the Ross “I’m FINE!” of quota management. Everyone can see it isn’t working.
One bit of good news. If you hit the limit mid-task, OpenAI says “Codex can continue working on that turn, subject to fair-use limits.” The block lands on the next turn, not halfway through a refactor.
When it shows up with usage left
This is the version that makes people angry. The three reports below are user reports, not confirmed bugs, and OpenAI hasn’t given a root cause:
- #57655 (Free plan, Linux, opened Oct 28, 2026): the reporter says Usage > Plan limits shows “100% remaining” and Analytics shows no recorded usage, yet Codex in the desktop app and the VS Code extension returns the message. Signing out, updating, restarting and clearing
~/.codex/auth.jsondidn’t help. A Plus user commented the same day with a five-hour limit message while both Codex and the web showed 100% left. - #36344 (Plus, CLI 0.146.0, opened Jul 31, 2026): the usage dashboard and
/statusboth showed 43% of the weekly limit remaining, but every new turn was rejected with “You’ve hit your usage limit” and a wait-until date about six days out. - #49773 (Pro 200, opened Oct 1, 2026): the reporter sees the “Add credits to continue” version only when running Codex over an SSH remote connection, not on fresh devboxes.
As of Oct 31, 2026, none of the three has an OpenAI staff reply in the thread. That’s frustrating. If you paid for a plan and your dashboard says you have usage left, “try again later” is a pretty thin explanation. It’s Lucille Bluth’s “I don’t understand the question, and I won’t respond to it,” as an error message.
What OpenAI does tell you to do here is specific. Support “does not reset ChatGPT or Codex usage limits”, but “If you believe your usage was counted incorrectly, or access has not returned after the stated reset time, contact Support so we can investigate.” Include “the limit or credit balance you’re questioning, the reset time shown, a screenshot, the Codex client and model, and when it happened—including your time zone” (help article, verified Oct 31, 2026). If you file a GitHub issue as well, codex doctor output is what the #36344 reporter attached. Strip anything sensitive first.
The Business and Enterprise version
On a work account the person holding the credit card isn’t you. OpenAI’s help center says:
- Business: workspaces can buy credits beyond included limits under “flexible pricing” (ChatGPT Business help, verified Oct 30, 2026). Premium seats ($125 a month, or $100 a month billed annually) get “higher usage for ChatGPT Work and Codex than Standard seats, no five-hour usage limit”.
- Enterprise on token-based billing in USD: “contact your workspace admin to check the workspace budget, your effective user limit, and the reset period shown for each control”, and “Follow your organization’s process for requesting additional usage.”
Google autocomplete suggests a workspace version of this message that points you to your workspace owner. That wording isn’t in any OpenAI help article or Codex issue we could cite, so it’s not quoted here. Either way, message your admin with your reset time. Whether Plus or a Business seat suits you better is in Codex Plus vs Business.
Ways to keep going (legit ones)
OpenAI lists the options as “adding credits, applying an available reset, upgrading, or waiting for the limit to reset” (help article, verified Oct 31, 2026). In the order I’d try them:
- Switch model for the next window. On Plus, GPT-6.1 Terra’s estimate is 35–285 messages per five hours against 15–160 on GPT-6.1 Sol, and GPT-6 Luna’s is 350–3,000 (Codex pricing, verified Oct 30, 2026). It won’t unblock you now, but the next window lasts longer.
- Wait for the reset. Free, and the dashboard tells you when.
- Use a banked reset, if you have one. Some Plus and Pro referral promotions add one. It shows in the profile menu as “1 reset available”, and using it “refreshes your 5-hour and weekly Codex usage windows and changes your weekly reset date.”
- Add credits. “If you hit a usage limit in Codex on Plus or Pro, you’ll see an option to add credits.” Buy them from Settings > Usage in ChatGPT or Usage & Billing in the Codex app. They’re “valid for 12 months from purchase”, non-refundable outside listed exceptions, and they’re not API credits. Free and Go: “Credit purchases are currently available to a limited group” (credits help, verified Oct 31, 2026). Per-credit price isn’t published in the article, so check current price at checkout. More in Codex credits.
- Use an API key. “All users may also run extra local chats using an API key, with usage charged at standard API rates” (Codex pricing, verified Oct 26, 2026). Good for CI and scripts. No cloud tasks.
- Upgrade. Plus to Pro gets rid of the five-hour window (“currently”), and Pro starts at $100 a month. Whether that’s worth it: Codex Pro vs Plus. Every tier and price: Codex pricing.
Skip extra free accounts, shared logins and “quota reset” scripts. They break the rules and can cost you the account that holds your work.
Not a quota problem? If the run died with “stream disconnected before completion”, see Codex stream disconnected. On Free and wondering what you get at all: Is Codex free?. Shopping for roomier limits: AI coding tools compared.
Here’s what I’d do
Plus, blocked mid-afternoon: run /status. If it’s the five-hour window, switch routine work to Terra or Luna and come back after the reset.
Plus, weekly limit gone: buy a small amount of credits to cover the rest of the week. If it happens every week, price out Pro.
Dashboard says you have usage left: screenshot it with the reset time, note your client, model and time zone, and send it to OpenAI Support. Add your case to the matching GitHub issue if it fits.
Business seat: ask your admin about workspace credits or a Premium seat.
Hitting stdin is not a terminal rather than a limit? The Codex CLI stdin fix covers pipes, CI and non-interactive mode.
Getting “Error submitting message” instead of a limit notice? This Codex error fix explains what each version means.
Not a limit message but turn/start failed in TUI? That one has its own causes: Codex turn/start failed.
A usage limit looks different from “Error loading messages”. Client load failures: Codex Error Loading Messages.
If the banner says the selected model is at capacity rather than that you hit a limit, that’s a different error: Codex selected model is at capacity.
FAQ
What does “You’ve hit your usage limit. Try again later.” mean in Codex? It means one of your plan’s Codex allowances, usually the five-hour window or the weekly limit, is used up until the reset time shown in Settings > Usage or /status.
When does Codex usage limit reset? OpenAI counts usage per five-hour period, weekly limits may also apply, and your exact reset time is shown in your usage dashboard or with /status in the CLI.
Does Codex Plus have a weekly limit? Yes, OpenAI says weekly limits may also apply on Plus on top of the five-hour estimates, but it doesn’t publish the weekly allowance size.
Can I buy more Codex usage? Yes, on Plus and Pro you can add credits from Settings > Usage in ChatGPT or Usage & Billing in the Codex app, and credits last 12 months from purchase.
Does starting a new chat fix the Codex usage limit error? No, OpenAI says starting a new chat does not reset your usage limits.
Sources
- openai/codex issue #57655 (Free plan): https://github.com/openai/codex/issues/57655 (verified 2026-10-30; re-read 2026-10-31)
- openai/codex issue #36344 (Plus): https://github.com/openai/codex/issues/36344 (verified 2026-10-30; re-read 2026-10-31)
- openai/codex issue #49773 (Pro 200, SSH): https://github.com/openai/codex/issues/49773 (verified 2026-10-30; re-read 2026-10-31)
- OpenAI Help Center, Using Codex with your ChatGPT plan (checking usage, options, Support policy): https://help.openai.com/en/articles/11369540-using-codex-with-your-chatgpt-plan (verified 2026-10-31)
- OpenAI Help Center, Using credits for flexible usage (buying, validity): https://help.openai.com/en/articles/12642688-using-credits-for-flexible-usage-in-chatgpt-personal-plans (verified 2026-10-31)
- OpenAI, Codex pricing (estimates, weekly wording, Pro rule): https://developers.openai.com/codex/pricing (verified 2026-10-30; API-key line verified 2026-10-26)
- OpenAI Help Center, What is ChatGPT Business (Premium seat, flexible pricing): https://help.openai.com/en/articles/8792828-what-is-chatgpt-business (verified 2026-10-30)
- OpenAI Help Center, About ChatGPT Pro plans (Pro 200 grandfathering ended Oct 29, 2026): https://help.openai.com/en/articles/9793128-about-chatgpt-pro-plans (verified 2026-10-29)