How ModelRush works

Live
One request contract, explicit routing, provider execution, and traceable records.
Open in ChatGPT
(opens in new tab)Last verified: 2026-08-11

Request lifecycle

  1. Your server authenticates with a ModelRush API key.
  2. The gateway validates the payload and model/endpoint compatibility.
  3. Region routing selects an eligible execution path.
  4. The provider executes the workload.
  5. ModelRush normalizes the response and records status, latency, errors, and billable usage.

Stable application boundary

Keep your product code coupled to the ModelRush endpoint contract, not to provider-specific SDKs. Provider differences remain visible where they affect parameters, availability, or data handling.

Traceability

Store response and generation IDs with your own user or job ID. Never place personal data in log labels merely to make requests searchable.
PreviousRegions and pricing
NextAsynchronous jobs