AI-103 Practice Question 53: Standard Deployment Quota Increase

TPM and RPM are the two rate-limit units Foundry enforces per model or deployment type, and Microsoft's quotas article gives one documented path to raise them: "To increase your quota, use Microsoft Foundry Service: Request for Quota Increase to submit your request." Requests are evaluated individually, and priority goes to customers already using their existing allocation, but the request form is the mechanism, not a workaround elsewhere in the architecture.

- A is wrong because a second deployment of the same model doesn't create independent capacity. For models under per-region quota management, TPM and RPM are capped per subscription and region, and the general best practice for a full quota page is to move quota from another deployment, not to conjure more of it by renaming a deployment. - B is wrong because shortening prompts reduces how much of the *existing* TPM budget each call consumes; it doesn't change the RPM number the deployment is allocated, and the developer asked for more headroom without changing the application. - D is wrong on the premise: Standard limits aren't frozen. The quota-increase request form explicitly covers Foundry Models sold by Azure and Azure OpenAI models, which includes Standard deployments; moving to PTU changes the pricing and SLA model but isn't required just to raise a rate limit.

Reference: Microsoft Foundry Models quotas and limits