Basic API access and shared inference.
Dedicated throughput and premium models.
Private clusters and custom AI deployments.