Question 1

Is cheapllm.ai compatible with the OpenAI API?

Accepted Answer

Yes. cheapllm.ai is designed around the OpenAI-compatible request format. In most cases, setup only requires changing base_url, api_key, and model.

Question 2

Do you store prompts or completions?

Accepted Answer

cheapllm.ai is designed with request privacy in mind. For production keys, request-body logging controls can keep only the metadata required for billing, abuse prevention, rate limiting, and reliability monitoring.

Question 3

How do you calculate token cost?

Accepted Answer

Each request is priced by model route: input tokens times input price plus output tokens times output price. The dashboard shows model, route, tokens, timestamp, and request-level cost.

Question 4

Are these the same models as official providers?

Accepted Answer

cheapllm.ai provides access to routes mapped to specific model IDs and versions. Route pages show model name, context window, pricing, availability, and last verified date.

Question 5

What happens if a route is unstable?

Accepted Answer

If a model route becomes degraded, the route page shows status updates. Depending on configuration, cheapllm.ai can retry, fail fast, or route eligible traffic to another available endpoint.