MODELS
Unlimited Basic subscription: unlimited Qwen 3.8 27B for €14.95/month, with no per-token billing

Documentation

Guides for deploying models, connecting the OpenAI-compatible API, and configuring routing, protection, and retrieval.

Inference API

Handle errors

Interpret authentication, policy, capacity and backend failures without losing the request trace.

Retry deliberately

Correct 4xx errors before retrying. Retry transient 5xx responses with exponential backoff and a bounded attempt count.

  • 401: missing or invalid API key.
  • 403: account policy or LLM Firewall decision.
  • 404: unknown serving name or resource.
  • 409: resource state does not allow the operation.
  • 429: request or capacity limit.

Keep the request id

Record the response request identifier with application logs and support reports. It connects the client failure to the QDivZero trace.