> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bettertoken.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Codex usage limit or API 429: which limit did you hit?

> Distinguish a Codex plan usage limit from an API rate limit, usage cap, or BetterToken balance issue before you retry.

A Codex usage warning and an API `429` are not the same limit. Identify the access mode first. This prevents repeated requests that cannot succeed.

## Identify the access mode

| What you use                                 | Where the limit belongs    | What to check                                                               |
| -------------------------------------------- | -------------------------- | --------------------------------------------------------------------------- |
| Codex with your official ChatGPT/Codex login | Your Codex plan or credits | The usage information shown by Codex and the official Codex usage guidance  |
| Codex with a BetterToken API Key             | Your API workflow          | HTTP status, response body, request ID, current balance, and request record |

<Warning>
  A BetterToken API Key does not extend or bypass an official Codex plan limit. It creates a separate API workflow with its own balance and request history.
</Warning>

## If Codex shows a plan usage warning

1. Save the current task state, changed files, and validation results.
2. Record the exact warning and the time it appeared.
3. Check the official [Codex usage limits and credits guidance](https://help.openai.com/en/articles/12642688).
4. If credits are available for your plan, follow the options shown in the official account UI. Otherwise, wait for the plan limit to reset.

Do not infer the remaining allowance from one percentage alone. Different UI indicators can describe different periods or resources.

## If the API returns 429

1. Save the HTTP status, response body, request ID, and any rate-limit or retry headers.
2. Check the current balance and the failed request in BetterToken Workspace.
3. Confirm that Codex uses the OpenAI-compatible Base URL `https://bettertoken.ai/v1` and your own BetterToken API Key.
4. Retry only when the response indicates a temporary rate limit. Use bounded exponential backoff and stop after a defined number of attempts.

If the response points to balance, access, or configuration, retrying the same request will not fix it. Correct that cause first, then send one small test request.

## Confirm the fix

Send a short, non-critical prompt. Confirm that it returns normally and that the request appears in the Dashboard with the expected model, status, Token usage, and charge.

For the complete setup, see [Connect Codex to BetterToken](/en/ai-tools/codex).

<Card title="Open BetterToken Workspace" href="https://bettertoken.ai/workspace?utm_source=docs&utm_medium=organic_content&utm_campaign=SEO-KW-203&utm_content=codex-limit-diagnostics">
  Check your balance, create your own API Key, and review request records.
</Card>

## Sources

* [OpenAI: Using Credits for Flexible Usage in ChatGPT (Free/Go/Plus/Pro)](https://help.openai.com/en/articles/12642688)
* [OpenAI Codex Help Center collection](https://help.openai.com/en/collections/14937394-codex)