Azure OpenAI and Azure AI Foundry

Organizations that want to use OpenAI's language models, such as the GPT models behind ChatGPT, in their own applications can do so via Azure. In 2023 Microsoft announced that ChatGPT was coming to the Azure OpenAI Service. Today that service is part of Azure AI Foundry. Here's what that means and how to start.

ChatGPT models have long been available in Azure. The Azure OpenAI Service now falls under Azure AI Foundry (the name may have changed slightly since). Which models and regions are available changes often; check the current documentation.

What is Azure OpenAI?

With Azure OpenAI you use OpenAI models via Microsoft's cloud, through an API. The difference with the regular ChatGPT website:

What is Azure AI Foundry?

Azure AI Foundry is Microsoft's platform for building AI applications and agents. It offers a catalog with models from OpenAI and other providers, tools to connect your own data, and features for evaluation and safety. You work via the portal or with the SDKs.

How to start

  1. Make sure you have an Azure subscription and the right permissions.
  2. Open the Azure AI Foundry portal and create a project or resource.
  3. Pick a model from the model catalog and deploy it in a region of your choice.
  4. Test your prompts in the playground.
  5. Use the endpoint and key, or better a managed identity, in your application.

Tips for IT

Timeline of Microsoft AI milestones up to 2023

Also see what is generative AI and secure score in Defender for Cloud.

Read also

What is generative AI and what can you do with it?
Get started with Microsoft 365 Copilot Chat (formerly Bing Chat Enterprise)
Secure score in Microsoft Defender for Cloud

Nederlandstalige versie: ChatGPT-modellen in Azure: Azure OpenAI en Azure AI Foundry


More in English