Check Codex usage limit: /status, /usage and the usage page

How to see Codex usage: the places that show it

OpenAI Codex usage dashboard

The pricing page points at the dashboard, and it recommends checking it every week or two to understand your pace and remaining capacity. It also says published message estimates are not fixed limits and that the dashboard is where current limits and reset times live. The same page says ChatGPT Work usage inside ChatGPT uses the same pricing, credits and usage limits as Codex, which answers the common “Chat GPT Codex usage” question: ChatGPT Work and Codex share the same usage limits. OpenAI’s Help Center article (Using Codex with your ChatGPT plan) gives the same route: open Settings or your usage dashboard, then check which allowance is exhausted, your credit balance and any reset time shown; and for current limits, rates and credit options, check the pricing page. Not found in the sources checked: whether the IDE extension has its own usage view.

Where the official sources disagree

Which URL is the usage page?

The pricing docs link https://chatgpt.com/codex/settings/usage. The open-source CLI prints a different address in its /status card (card.rs) and in its limit-reached messages (error.rs): https://chatgpt.com/settings/usage. Neither could be opened without signing in, so it is not verified whether they are the same page. That is a string in the open-source CLI, not documentation.

Does /status show limits, or does /usage?

The pricing page says /status shows your remaining limits. The CLI command reference lists only the active model, approval policy, writable roots and current token usage under /status, and assigns account usage to /usage. The open-source CLI's own test snapshots show /status rendering limit rows, so in practice it does; the documentation just describes it inconsistently.

What the /status limit rows look like

The CLI repository keeps test snapshots of the /status output. Per those snapshots, the card has a Context window row and then 5h limit, Weekly limit and Credits rows, each limit shown as a bar followed by NN% left (resets HH:MM); a separate snapshot has a Monthly limit row whose reset text includes a date. The lines below are copied from two of those snapshots (credits and limits, monthly limit); the figures are the test's mock values, not anyone's real usage and not a statement of any plan's limits.

5h limit:        [███████████░░░░░░░░░] 55% left (resets 09:25)
Weekly limit:    [██████████████░░░░░░] 70% left (resets 09:55)
Monthly limit:   [██████████████████░░] 88% left (resets 07:08 on 7 May)

Paste your own output into the parser on the usage-limits page to read it more easily; it runs in your browser and sends nothing anywhere. Two cautions from the source code, not the docs: the CLI can warn limits may be stale - run /status again shortly., and the reset times appear to be formatted in the machine's local time zone, but no docs page states which time zone /status uses.

What the CLI prints when a limit is hit

These are strings in the open-source CLI's error.rs — not documentation. They all start “You’ve hit your usage limit.” and then differ by plan: the Free/Go variant says “Upgrade to Plus to continue using Codex”, the Pro variant says “Visit https://chatgpt.com/settings/usage to purchase more credits”, and the Business/Team variant says “To get more access now, send a request to your admin”. The ending is either a specific retry time or “try again later”. Per the pricing docs, a turn already in progress when you reach the limit can keep going.

For developers: the app-server rate-limit fields

The app-server protocol docs describe an account/rateLimits/read method. Each limit window it returns carries usedPercent (current usage within the window), windowDurationMins (the window length) and resetsAt (a Unix timestamp, in seconds, for the next reset), per the app-server docs.

Not found in the sources checked

• Which time zone the reset times in /status use.
• A complete description of every /status row (the docs name only model, approval policy, writable roots and token usage; the row layout above comes from repository snapshots).
• Whether the IDE extension has a usage view.
• Whether the two usage-page addresses are the same page.
• Whether the five-hour window is rolling or fixed, and when exactly it starts; and what day or time a weekly window resets (the docs and the Help Center article say to check the dashboard).

For what a reset does and does not change, see how to reset a Codex usage limit; for how announced resets are tracked here, see the home page and usage limits explained.