Troubleshooting

Here are the most useful checks when a request isn't working.

I can't send a message

Check that a model is selected and that your message contains text or a file. On the free tier, respect the message size limit and pick a model accessible on that tier.

Error 401: incorrect API key

The key is missing, invalid, or revoked. Create a new key in Settings → API keys, update your environment variable, then restart your application.

Error 402: insufficient credits

Your balance is exhausted. Top up your account or wait for your next billing renewal.

Error 403: subscription required

The selected model or API usage requires an active subscription. Choose a plan or use a model available on the free tier.

Error 404: model not found

Check the identifier sent to the API. Copy the technical identifier from the Models page; the public name shown in chat can differ.

The model gives an unexpected answer

Clarify the context, expected format, and constraints. Try another model category: a reasoning or code model may fit better than a fast model depending on the task.

Need more help

Prepare the error message, the model identifier, the approximate time, and the minimal context needed to reproduce the issue. Never share your API key in a support request.