Inference Endpoints are billed per minute while initializing or running
Catalog row in Vendor License Rules · Cited
- Kind
- Counting
- Statement
- Inference Endpoints pricing documentation: charges apply for compute used while a deployed Endpoint is initializing or running; prices are shown per hour but the cost is calculated by the minute; access requires an active subscription and a credit card on file.
- Applies when
- Deploying a dedicated Endpoint
- Applies to
- Inference Endpoints
- Related metrics