Plan and manage an Azure AI solution 1.1 Choose the appropriate Foundry services for generative AI and agents

A procurement policy allows only models that are billed through the company's Azure…

Study guide skill: Choose an appropriate model for each task, including large language models (LLMs), small language models, multimodal models, and Foundry Tools

Question B6-001

A procurement policy allows only models that are billed through the company's Azure subscription, covered by Azure service-level agreements, and supported by Microsoft. An architect is shortlisting models from the Foundry model catalog. Which two statements are correct? (Choose two.)

Select 2 answers.

  • A Models sold by Azure are deployed by using serverless API deployments ✓ Correct answer
  • B Treat every catalog model the same way, because all models in the catalog are sold by Azure
  • C Shortlist only Azure OpenAI models, because no other provider's models are sold by Azure
  • D Deploy the shortlisted models on managed compute (preview), the option Foundry uses for models sold by Azure
  • E Shortlist from Foundry Models sold by Azure, which include all Azure OpenAI models and selected models from other top providers ✓ Correct answer

Answer: A, E

Foundry Models sold by Azure are billed through your Azure subscription, covered by Azure SLAs, and supported by Microsoft, and they include all Azure OpenAI models plus selected models from top providers. They use serverless API deployments. The catalog also has a second category, models from partners and community, and managed compute (preview) is for open-source, partner, and custom models on dedicated GPU capacity.

Deep dive

How it works

The Foundry model catalog has two main categories:

  • Foundry Models sold by Azure: hosted and operated by Azure. They include all Azure OpenAI models and specific, selected models from top providers. They're billed through your Azure subscription, covered by Azure service-level agreements and supported by Microsoft.
  • Foundry Models from partners and community: everything else in the catalog, with different terms.

The procurement policy restates the definition of "sold by Azure", so that's the list to shortlist from, and it isn't limited to OpenAI. The models page has tabs for Azure OpenAI and for other collections from providers like Cohere, DeepSeek, Meta, Mistral AI and xAI.

For deployment, Foundry has two options:

  • Serverless API: for Foundry Models, including those sold by Azure and select partner and community models. It's the preferred, most capable path, with standard, provisioned throughput, batch and developer deployment types.
  • Managed compute (preview): for open-source, partner and custom models on dedicated GPU capacity that Foundry manages. As a preview it comes without an SLA, which conflicts with this policy anyway.

The other options, and when you'd use them

  • Treat every catalog model the same: The catalog has two categories, and only one is sold by Azure.
  • Only Azure OpenAI models: Models sold by Azure also include selected models from other top providers.
  • Managed compute (preview): It's for open-source, partner and custom models. Models sold by Azure use serverless API deployments.

Exam tip

Sold by Azure = Azure billing + Azure SLA + Microsoft support, all Azure OpenAI plus selected others, deployed as serverless API.

References

Last reviewed:

Ready to test yourself? The quiz asks this and every other AI-103 question, with instant feedback. Practice this in a quiz →

All Plan and manage questions · AI-103 study guide

Thanks for your feedback!