Pass IndexThe State of AISign in

Provisioned throughput

Renting a fixed slice of a seller's hardware by the hour or the month instead of paying per token.

how it is sold · the vocabulary

reserved capacitycommitted throughput

It buys guaranteed capacity and predictable latency, and it is billed whether you use it or not. It becomes cheaper than per-token pricing only at steady, high volume — and the break-even is usually much higher than people expect.

Nearby

API keyBatchCommitted spendCreditDeprecationFree tierFrontier modelGateway