A provider-run inference service billed per token, where the provider — not you — owns the runtime, scaling, and cold starts.
Continue to AI University →