Try a AI-103 question.

Plan and manage an Azure AI solution

An engineer is migrating a Provisioned deployment from an older model to its recommended replacement as part of routine model-lifecycle maintenance. The replacement model requires more PTUs per unit of throughput than the old one. What must the engineer do when configuring the new deployment?

• ARecalculate and provision the PTUs the replacement model needs
• BReuse the exact PTU count from the old deployment, since PTU capacity is model-agnostic
• CSwitch the deployment to a Standard type instead, since Provisioned deployments cannot change models
• DRequest a subscription-level token-per-minute quota increase to cover the new model's throughput needs