InferenceHelp

Error codes

Which error codes the Yenisei API returns, what each of them means, and what to do about it.

If a request fails, the API responds with an HTTP code and a body in the OpenAI format: the message field holds a description in words, and the code field holds a short code that makes the error easy to handle in a program. This page lists all the codes the API returns and what to do about each.

JSON

{
  "error": {
    "message": "Model not available on your plan.",
    "type": "invalid_request_error",
    "code": "model_not_allowed"
  }
}

Response codes

401
invalid_api_key: the API key is missing or invalid. Check the Authorization header and the key itself in the console.
403
model_not_allowed: the model is not included in your plan. Choose another model or change your plan.
404
model_not_found: there is no such model in the catalog. The list of available models is returned by the GET /v1/models method.
429
session_limit_exceeded, week_limit_exceeded or month_limit_exceeded: the credit limit in one of the windows has been exhausted. The Retry-After and X-RateLimit-Reset headers say when the limit resets. See the Limits and usage page for details.
502
bad_gateway: the model did not respond. Retry the request.

Health check

The GET /healthz method responds with {"status":"ok"} and does not require a key. It is a convenient way to check whether the API is available.

Back

Reasoning

Limits and usage