AI-300 Practice Question 70: Batch Deployment Cost

Microsoft Machine Learning Operations Engineer Associate (AI-300) · Test 7 · question 10 of 10

Optimize generative AI systems and model performanceeasy

A company classifies 2 million support tickets every night. Results are needed within 24 hours, and the team wants the lowest cost without consuming the quota used by its interactive chat application. Which approach should they use?

Select one answer
Show answer and explanation

Answer:

  • A. A Global Batch deployment with a batch job

The Batch API processes large groups of requests asynchronously, with separate quota and a 24-hour target turnaround, at 50 percent lower cost than Global Standard. The nightly job tolerates the delay, so a Global Batch deployment fits (option A).

- B is wrong because this is billed at the regular pay-per-token rate and shares quota with the interactive application. - C is wrong because reserved capacity is billed all day and is intended for steady, latency-sensitive traffic, so it costs the most here. - D is wrong because the Developer type is intended for evaluating fine-tuned candidates, has no SLA and is not a bulk processing option.

Reference: Getting started with Azure OpenAI batch deployments

Practice Test 7 ›

Study guides for AI-300