Your team is deploying a customer-facing chat application on Azure AI Foundry. The application experiences predictable, sustained high traffic during business hours with strict latency SLAs, and finance requires cost predictability rather than variable per-token billing. Which model deployment option should you choose to meet these requirements?