Errors

Error shape

OpenAI-format endpoints (/v1/chat/completions, /v1/responses, /v1/models):

{
  "error": {
    "message": "Model 'nope/nope' not found (request id: req_01ba...)",
    "type": "invalid_request_error",
    "code": "model_not_found",
    "request_id": "req_01ba..."
  }
}

Anthropic-format endpoint (/v1/messages):

{
  "type": "error",
  "error": {
    "type": "not_found_error",
    "message": "Model 'nope' not found (request id: req_2062...)",
    "request_id": "req_2062..."
  }
}

Include the request_id when you contact support. With it we can trace your request.

HTTP status

Status	Example message	Meaning	What to do
400	model is required, messages is required	The request is incomplete or malformed	Fix the request body
400	Model '...' matches several models (ambiguous_model)	A short name matches several providers	Use one of the full ids in the message
401	API key is required, Invalid API key	No key sent, or a wrong key	Check the header and copy the key again
402	Insufficient balance	The balance does not cover this request’s estimated cost	Top up, or lower max_tokens
403	API key is inactive	The key is turned off	Turn it on again in API Keys
403	This account has been blocked (account_blocked)	The account is blocked	Contact support
404	Model '...' not found (model_not_found)	Wrong model id, or the model is inactive	Get the id from GET /v1/models
429	Passed on from the model	The target model is rate limiting	Wait a moment and retry
502, 503	The model's supplier is currently unavailable	The model has a problem on the provider side	Retry shortly, or use another model. Vikey Auto moves to the next level automatically.

Handling errors in your app

• Retry 429, 502 and 503 automatically, with growing delays.
• Do not retry 400, 401, 402, 403 and 404 automatically. The result stays the same until the cause is fixed.
• For 402 in an app used by many people, show a clear message to the app’s admin so the balance is topped up quickly.