Skip to content

Common Error Codes

When you encounter an error, first check the returned error code, then refer to the table below for troubleshooting.


Status CodeDescriptionSolution
400Invalid request formatCheck the request parameters; for example, o1 series models do not support the system parameter
401Invalid tokenCheck whether the API key is correct; you can test with another model for verification
403Token group disabledEdit the token to remove the quota limit, or create a new token
404Endpoint does not existCheck whether the Base URL is correct; try adding /v1 or a trailing slash /
413Request content too longShorten the prompt and try again
429Upstream rate limitThe account has too much concurrency; try again later
500Internal server errorIf repeated retries still fail, contact the administrator
503Model unavailableThe current group does not have a channel for this model. Check whether the model name is correct, and whether there are extra or missing spaces, etc.
504Gateway timeoutThe upstream server did not respond in time; try again later
524Connection timeoutThe channel is congested; try again later

400 error code usually means the request parameters do not match. Comment out system first and see whether it works.

401 error code, “Invalid token,” is caused by a mismatch between the token and the API address URL.

403 error code usually means the token quota is insufficient (token quota and account quota are two different concepts). Go to token editing, click “Set to unlimited quota,” and it should usually take effect in about 2 minutes. You can also create a new token and set it to “unlimited quota” to try again.

404 error code is most likely because the API address was not set correctly.

429 error code indicates that the model has hit its traffic limit. In simple terms, too many users are causing the model to reach TPM saturation. In most cases, you can use it again after waiting a while and retrying. If it still does not work, send the model name to customer support and they will investigate.

503 error code: please check whether the full model name is correct. You can find the full model name on the top menu under the Supported Models page. If the model name is correct, send it to customer support for investigation.


4All API - One-stop AI large model API aggregation platform
Official website: https://4allapi.com
API Base: https://api.4allapi.com