Codex "You've hit your usage limit"

What does the message mean?

Your plan’s Codex allowance for the current window is gone. OpenAI’s pricing page says Plus and Standard Business plans get an estimated number of local messages per five hours, and weekly limits may also apply. Pro plans currently have no five-hour limit.

The wording after the first sentence depends on your plan. Codex is open source, and these are the messages in its code (codex-rs/protocol/src/error.rs):

Plan	Message
Plus	You’ve hit your usage limit. Upgrade to Pro (https://chatgpt.com/explore/pro), visit https://chatgpt.com/settings/usage to purchase more credits or try again at <time>.
Pro	You’ve hit your usage limit. Visit https://chatgpt.com/settings/usage to purchase more credits or try again at <time>.
Business and Team	You’ve hit your usage limit. To get more access now, send a request to your admin or try again at <time>.
Free and Go	You’ve hit your usage limit. Upgrade to Plus to continue using Codex (https://chatgpt.com/explore/plus), or try again at <time>.
Enterprise and Edu	You’ve hit your usage limit. Try again at <time>.

When Codex doesn’t know the reset time, the ending reads “try again later” instead.

Workspaces that pay with credits get different messages. Your workspace is out of credits. Add credits to continue. goes to an owner. Members see Ask your workspace owner to refill in order to continue. And You hit your spend cap set in your workspace. means a spend cap stopped you, not the plan.

How do you keep working?

1. Check the reset time. It’s at the end of the message, and on the usage dashboard. In the CLI, run /status.
2. Buy credits if you’re on Plus or Pro. OpenAI says you can purchase more and keep going without changing plans.
3. On Business, Edu or Enterprise with flexible pricing, the workspace can buy more credits. If you’re not the admin, send the request the message mentions.
4. Use an API key. Any user can run extra local chats with one, billed at standard API rates.
5. Or wait for the reset.

If the limit hits while Codex is mid-turn, OpenAI says it can finish that turn, subject to fair use limits.

How do you hit it less?

OpenAI’s advice is to look at the dashboard every week or two to learn your pace. Model choice, context size and tool use all change how fast a task eats the allowance. A smaller model or a tighter task often gets the same result for less.

Credits and API usage both land on a card. ernd is a credit card for US companies with 4% back on AI vendor spend and 1.5% on everything else. Request the card at app.ernd.ai.