Error codes

Every error has the same shape, compatible with the OpenAI SDKs. Always log the request_id; it’s also in the x-request-id header.

json

{
  "error": {
    "type": "invalid_request_error",
    "code": "deployment_not_found",
    "message": "No deployment named 'contract-reveiw' in this organization.",
    "request_id": "req_01a1843c3ca17daf8cc7837ff89e46b3"
  }
}

Inference API codes

Code	HTTP	Meaning	What to do
`invalid_api_key`	401	The key is missing, malformed or unknown.	Check the Authorization header and the key value.
`key_revoked`	401	The key was revoked.	Create a new key.
`key_expired`	401	The key passed its expiry date.	Create a new key.
`deployment_not_found`	404	No deployment with that alias in your organization.	Check the model value against your deployment aliases.
`deployment_paused`	409	The deployment is paused.	Resume it in the console.
`deployment_not_allowed_for_key`	403	The key is limited to other deployments or another client.	Use a key allowed for this deployment.
`unsupported_parameter`	400	The request uses a field Basevex doesn’t support yet.	Remove the field; see the Inference API reference.
`context_length_exceeded`	400	Input plus max_tokens exceeds the model’s context length.	Shorten the input or lower max_tokens.
`rate_limited`	429	Too many requests for this deployment or organization.	Retry with exponential backoff.
`spend_cap_reached`	402	The request could push spend over your monthly cap.	Raise the cap or wait for the next month.
`subscription_inactive`	402	Your subscription isn’t active (for example, a failed payment).	Update your payment method in Billing.
`org_suspended`	403	The organization is suspended.	Contact support.
`ip_not_allowed`	403	The API key has an IP allowlist that doesn’t include your address.	Call from an allowed address or update the key’s allowlist.
`runtime_unavailable`	503	Inference capacity is temporarily unavailable.	Retry with backoff.
`runtime_timeout`	504	The model took too long to respond.	Retry, or lower max_tokens.
`internal_error`	500	An unexpected error on our side.	Retry; contact support with the request_id if it persists.

Retrying

Retry 429, 500, 503 and 504 with exponential backoff and jitter. Don’t retry 4xx errors other than 429 without changing the request.