Error messages

You exceeded your current quota, please check your plan and billing details.

You exceeded your current quota, please check your plan and billing details.

This is the OpenAI API's refusal when your account can't pay for the request. The full text, as a captured error in OpenAI's own Codex repo has it:

You exceeded your current quota, please check your plan and billing details. For more information on this error, read the docs: https://platform.openai.com/docs/guides/error-codes/api-errors.

It arrives as HTTP 429 with the type insufficient_quota. It's also old and very common: 4,751 GitHub issues contain insufficient_quota as of September 23, 2026, and it's still being served. Someone's coding agent logged it on August 26 against gpt-5.6-sol, word for word.

A 429 that isn't about speed

429 usually means "slow down", which is why so many people wait and retry. This one is about money. OpenAI's error-code page is plain about it: "Retrying billing, spend, or quota errors won't restore API access." The Python SDK doesn't know the difference, though. In openai-python 3.19.0 every 429 becomes a RateLimitError and is retried twice by default (unless the server's x-should-retry header says no), so your code may pause a few times before the error reaches you.

The four codes behind it

The surprise when we read the current error-code page was that this sentence isn't on it any more. OpenAI now documents four separate 429 codes and says to read error.code, adding that "the broader error.type can still be insufficient_quota":

• credit_balance_exhausted: the organization's prepaid credits are gone. Add credits in billing settings.
• organization_spend_limit_exceeded: you hit the monthly spend limit you set for the whole organization. Raise or remove it, or wait for the monthly reset.
• project_spend_limit_exceeded: the same, for one project. Other projects keep working unless their own limit is also hit.
• organization_usage_limit_exceeded: OpenAI's own monthly cap for your organization, separate from anything you set. You ask OpenAI for a higher one.

So if your error body carries a code, it tells you which page of the dashboard to open. If it only has the old sentence, check the billing page and the limits page both.

In Codex

Codex turns all of these into one shorter line of its own:

Quota exceeded. Check your plan and billing details.

That string is in the 0.156.1 release. Codex doesn't retry it, since it's in the same terminal group as its usage-limit errors. Codex has its own message for a spent ChatGPT plan, so the quota line is a sign Codex was working from an API account.

That tripped people up in 2025. An August report on Codex 0.14.0 had someone signed in with ChatGPT and still spending API credits, because an OPENAI_API_KEY was set in their environment. A Codex maintainer suggested starting it as OPENAI_API_KEY= codex to blank the key, and said the underlying bug was fixed from 0.19.0. If you meant to use your ChatGPT plan and see this line, check for a stray key first.

Not to be confused with

Google's Gemini API uses almost the same opening words, "You exceeded your current quota, please check your plan and billing details.", but with RESOURCE_EXHAUSTED and a link to ai.google.dev. There it usually is a rate limit, so the fix is different: the Gemini page covers it.