One request. One model. One market behind it.
Infera is one API for the leading frontier models, routed to the best price available for each request and settled request by request.
Only models that are live and healthy can receive requests. You pick the model, Infera finds capacity for it, and payment runs through your prepaid Infera balance.
One interface for frontier models
Every supported model is reachable through the same API. The catalog shows what each model can do, what it costs by default and whether it can take traffic right now.
Routed, never substituted
Infera routes a request only for the model you named. A different model is never swapped in, even when it would be cheaper or faster.
Markets behind the request
Behind each call sits a market: teams buying inference on one side, compute supply on the other, and a clearing layer that prices the request as it arrives and settles it on measured usage. Each request keeps a receipt with token counts, the applied rate and the final charge. Prompts and completions are not stored.