How do Codex usage limits work with ChatGPT?
Codex draws on your ChatGPT plan, over a 5-hour window, sometimes with a weekly limit. Here is the gist, based on OpenAI’s documentation.
The 5-hour window
OpenAI gives message estimates “per five-hour period”, adding that they “are not fixed message limits”. Local messages and cloud tasks share the plan’s allowance, and “weekly limits may also apply”.
| Plan | Price |
|---|---|
| Go | $8 a month |
| Plus | $20 a month |
| Pro | From $100 a month, for 5x or 20x Plus’s limits |
| Business | $25 per user a month, or $20 billed yearly |
US prices shown by OpenAI on September 29, 2026; they vary by country.
See what you have left
In the Codex CLI, the /status command shows your usage. Codex’s usage page, in your ChatGPT settings, shows it too.
At the ceiling
If you hit the limit mid-task, “the agent will be able to continue working on that turn”. After that, you can wait for the displayed reset time, move to a higher plan, or buy credits: the plan’s allowance is used first, then credits take over. They stay valid for 12 months.
Frequently asked questions
Do Codex and ChatGPT share the same limits?
Largely yes: OpenAI says ChatGPT work usage inside ChatGPT draws on the same limits as Codex.
How many Codex messages per 5 hours?
OpenAI gives ranges by plan and model, noting they are not fixed limits: they depend on the size of your tasks.
Where can I see my Codex usage?
With /status in the CLI, on Codex’s usage page, or in JaugeAI next to your Claude limits.