How to check your Codex usage limits and when they reset.
Codex counts your work against your ChatGPT plan. This guide shows where to see your limits, how to keep them on screen while you work, and what OpenAI's own pages say about resets.
Type /status in a running Codex CLI session to see your remaining usage limits, as OpenAI's pricing page says. For current limits and reset times, open the usage dashboard at chatgpt.com/codex/settings/usage. On Plus and Standard Business plans, usage is counted per five-hour period, and weekly limits may also apply. Pro plans currently have no five-hour limit.
Cowork Island
macOS 14 or later.
Apple silicon and Intel.
No account needed for Free.
Where to see your Codex usage limits.
OpenAI documents four places that show your Codex limits, plus one command that is easy to confuse with them. OpenAI's pricing page also says ChatGPT Work and Codex share usage, so your ChatGPT Work usage counts toward the same limits.
| Place | How to open it | What it shows |
|---|---|---|
| Codex CLI | Type /status in a running session. | Your remaining limits, according to the pricing page. |
| Codex CLI footer | Type /statusline and turn on rate limits. | Your limits at the bottom of the screen, kept after a restart. |
| ChatGPT desktop app (Codex) | Type /status in the chat composer. | The chat ID, context usage and rate limits. |
| Usage dashboard | Open chatgpt.com/codex/settings/usage in a browser. | Your current limits and reset times. |
Codex CLI, /usage | Type /usage in a running session. | Account token activity, or an earned reset to redeem. The docs do not list your limits here. |
The ChatGPT desktop app is where the old Codex app now lives. OpenAI's changelog for 9 July 2026 says that Codex “is now part of the ChatGPT desktop app.” In that app, /usage opens usage and billing settings, when available with ChatGPT sign-in.
Use /status for limits, not /usage.
OpenAI's pages describe these two commands in different words, so it is easy to pick the wrong one.
/statusis the one for limits. The pricing page says: “If you want to see your remaining limits during an active Codex CLI session, you can use /status.” The CLI command reference describes it only as “session configuration and token usage” and leaves limits out. The two pages do not contradict each other./usageis for activity. The command reference says it lets you “View account token usage or use a rate-limit reset.” It opens a menu with your token activity (/usage daily,/usage weeklyor/usage cumulative) and, when you have one, an earned reset to redeem.
/status
The docs do not describe what the /status screen looks like. A test snapshot in Codex's public source code shows a “5h limit” line and a “Weekly limit” line, each with the percentage left and the reset time. This comes from source code, not the docs, so the layout can change from one release to the next.
Keep your limits in view with /statusline.
If you check often, pin your limits to the footer, the line at the bottom of the Codex CLI screen. OpenAI's command reference says /statusline lets you pick and reorder footer items. Rate limits are one of them; others include context stats, the git branch and token counters.
- In a running Codex CLI session, type
/statuslineand press Return. - In the picker, turn on rate limits. Reorder the items if you like, then confirm.
- The footer updates at once. Codex saves your choice as
tui.status_linein itsconfig.toml, so it is still there after a restart.
Let Codex write this setting for you. If you later edit config.toml by hand, merge your change into the file instead of replacing it. If a [tui] section is already there, add keys under it, and never add a second [tui] header.
How Codex usage limits reset.
The OpenAI pages we could read do not publish a fixed reset time for Codex. The pricing page sends you to your own numbers instead: “check your usage dashboard for current limits and reset times.” Go by the reset time on your dashboard.
The pricing page does say this about how fast you use your allowance:
- The per-model figures are estimates of local messages per five-hour period. OpenAI says they “are not fixed message limits.”
- Usage depends on the model and the task. Cloud tasks may use more of your allowance than local messages, and both share the same allowance.
- Fast and Ultrafast speed modes use your included limits at a higher rate than Standard mode for the same model.
- ChatGPT Work inside ChatGPT uses the same usage limits as Codex.
If you have an earned reset, /usage offers to redeem it. OpenAI's help center has articles on how the five-hour and weekly windows work. They did not load when we checked on 18 October 2026, so this guide does not repeat their rules. If you read them, compare what they say with your dashboard.
Five-hour and weekly limits by plan.
Codex is included in ChatGPT Free, Go, Plus, Pro, Business, Edu and Enterprise plans. The pricing page describes the limits like this:
| Plan | Five-hour limit | Weekly limit |
|---|---|---|
| Plus, Standard Business | Yes. The page gives estimates per model, not fixed numbers. | May also apply. |
| Pro | None at present: “Pro plans currently have no five-hour limit.” | May also apply. |
| Enterprise and Edu with flexible pricing | No fixed rate limits. Usage scales with credits. | |
| Enterprise and Edu without flexible pricing | The same per-seat limits as Plus for most features. | |
The five-hour estimates on the page cover Plus and Standard Business only. For any other plan, check your dashboard. The figures change often, so this guide does not copy them.
What happens when you reach the limit.
A turn that is already running can finish. OpenAI says: “If you reach your usage limits during an active turn, the agent will be able to continue working on that turn, subject to fair use limits.” After that, you can wait for the reset time shown on your dashboard, or use one of the ways the pricing page lists:
- On ChatGPT Plus or Pro, buy additional credits without upgrading your plan. Business, Edu and Enterprise plans with flexible pricing can buy workspace credits.
- On any plan, run extra local chats with an API key, charged at standard API rates.
The no-setup option.
Cowork Island is a macOS 14+ app that keeps your AI agent sessions from supported tools in a notch panel, showing the state of each one. You can follow Codex, Claude Code and more in one place and see which session “Needs you”. By default, the island opens by itself when a session starts needing you, unless that session's terminal or app is already in front.
Its usage meters show your own Codex and Claude plan limits, five-hour and weekly. For Codex, the figures come from Codex's own session files on your Mac, so there is nothing extra to turn on for them. Claude's figures need a one-time step in Settings and a Claude Pro or Max plan. To see Codex sessions themselves, switch Codex on in Settings and allow Cowork Island's hooks in Codex.
Free keeps working with one session card at a time and a one-line usage summary in the panel header. Pro shows all your sessions, lets you jump back to them and adds detailed usage meters, with reset times, for each account. A 7-day Pro trial starts by itself, once per Mac, without payment details.
Download for freeSources.
- OpenAI: Codex pricing and usage limits
- OpenAI: Codex CLI commands
- OpenAI: slash commands in the ChatGPT desktop app
- OpenAI: Codex changelog
- openai/codex on GitHub (source code, not documentation)
Sources checked 18 October 2026. This guide is written by Cowork Island's maker. For corrections, email support@coworkisland.com or open an issue on GitHub.