(1) Charged in increments of 5,000 tokens per judge request. See illustrative sizing example below for details.

* For regional availability: AWS, Azure, GCP, SAP