Codex Usage Limits: Check Usage and Reset Times
To check your Codex usage limits, open the Codex usage dashboard.
In an active Codex CLI session, enter /status to see your remaining allowance. Use the reset times shown for your account when deciding when to resume work.
If you’ve reached a limit, first identify which allowance is exhausted. A short-term limit, a weekly allowance, and API billing require different next steps.
Open the usage dashboard
Open the usage page in your browser and sign in with the account you use for Codex.
Read the remaining allowance and reset time together. If you see multiple limits, check each one before planning another session.
For example, imagine your dashboard shows:
| Allowance | Remaining | Reset |
|---|---|---|
| Short-term | 0% | Today at 3:00 PM |
| Weekly | 40% | Friday at 10:00 AM |
This is an illustrative example, not a screenshot or a universal schedule. The exhausted short-term allowance is the immediate problem. If the weekly allowance were also exhausted, the earlier reset wouldn’t tell the whole story.
Pay attention to whether the interface labels a percentage as used or remaining. “90% used” and “90% remaining” mean very different things.
Check usage in the terminal
Inside an active Codex CLI conversation, enter:
/status
This is a slash command for the Codex prompt. Don’t enter it directly into your normal shell.
For account token activity, recent CLI versions also provide:
/usage
The usage menu lets you view token activity or redeem an available earned reset. You can open specific activity views directly:
/usage daily
/usage weekly
/usage cumulative
These commands show activity for the selected period. A cumulative token total doesn’t tell you the remaining percentage of your subscription allowance.
Check usage while working in VS Code
You can use the browser dashboard while working in the extension. Confirm that the browser and extension are signed in to the same account and workspace.
The extension also documents /status as a command for viewing context usage and rate limits. Type / in the Codex composer to see the commands available in your installed version.
If you’re still setting up the extension, start with how to use Codex in VS Code.
How five-hour and weekly limits work
OpenAI currently describes five-hour usage estimates for Plus and Standard Business. Pro plans currently have no five-hour limit. Weekly limits may apply, and local messages and cloud chats share the plan allowance. The published message ranges are estimates, not fixed message quotas.
That makes “How many prompts do I get?” a less useful planning question than “How much allowance do my usual tasks consume?”
Try recording the remaining percentage before and after a typical task. For example, compare a small bug fix with a larger feature request over several sessions. Use those observations to plan your workload; don’t treat one task as a reliable cost estimate for every future task.
When do Codex limits reset?
Use the reset timestamp shown in your dashboard. Don’t assume that another user’s schedule, midnight, or your subscription billing date is your reset time.
If your account shows two allowances, record both timestamps. For example:
Short-term allowance resets: today at 15:00
Weekly allowance resets: Friday at 10:00
When checking again later, refresh the page and read the current values. A saved screenshot is useful for comparison, but it isn’t a live usage meter.
Also distinguish a scheduled reset from a redeemable reset. The CLI’s /usage menu can offer an earned reset when one is available. Read the terms shown before redeeming it; the presence of the command doesn’t mean every account has a reset to spend.
What to do when you reach a usage limit
Check the dashboard before retrying the same request. OpenAI documents additional credits for eligible plans and API-key billing as ways to continue local work. An active turn may continue after reaching a limit, subject to fair-use limits.
Choose your next step based on the work:
- If the task can wait: note the displayed reset time and return later.
- If you’re considering credits: check the offered price and how much work remains before buying.
- If an earned reset is available: review the redemption details before using it.
- If this happens regularly: compare your typical workload with the available plans rather than upgrading because of one unusually large task.
You can also use the waiting time to review completed changes, run tests yourself, and prepare a more precise next request.
Why token usage isn’t your remaining quota
There are several numbers you may encounter:
| Number | What to use it for |
|---|---|
| Remaining allowance and reset time | Deciding whether you can continue with included usage |
| Token activity | Understanding how much text and other context your work processes |
| Context usage | Understanding the current conversation’s size |
| API charges | Tracking separately billed API usage |
For example, seeing a large cumulative token count doesn’t prove that you’ve exhausted this week’s allowance. Likewise, a conversation with room left in its context can still encounter an account usage limit.
Use the meter that answers your question: remaining allowance for availability, token activity for activity, and billing information for spend.
What if you’re using an API key?
Codex supports both ChatGPT sign-in and API-key authentication. API-key usage follows OpenAI Platform billing rather than your included ChatGPT plan credits.
To check your CLI authentication method, run this in your normal terminal:
codex login status
Unlike /status, this is a shell command. It checks authentication; it doesn’t show your remaining subscription allowance.
If you’re using an API key, inspect your Platform usage and the exact error message. Waiting for a ChatGPT allowance to reset won’t resolve an unrelated API billing or rate-limit issue.
Frequently asked questions
Does opening a new chat reset my Codex allowance?
Treat a new chat as a fresh conversation, not evidence that your account allowance has renewed. Check the usage dashboard before assuming more quota is available.
Why doesn’t my account show a five-hour limit?
The available meters depend on your plan. In particular, OpenAI currently states that Pro plans have no five-hour limit. That doesn’t establish unlimited usage.
Why do the dashboard and CLI show different numbers?
First check that you’re comparing the same account, workspace, and metric. Token totals, context percentages, and remaining allowance aren’t interchangeable. Refresh the dashboard and run /status again.
If the same allowance still disagrees, record the two readings and the time you checked them so you can describe the discrepancy clearly when requesting support.
What if the reset time has passed but I’m still blocked?
Refresh the usage page and check every displayed allowance. Read the new error message rather than assuming the old limit is still the cause.
If an allowance appears available but the request fails, note the error, account workspace, client version, and time. Those details help distinguish a usage problem from a sign-in or connection problem.