Use of OpenAI models in Azure, including GPT, o-series, embeddings, image and Whisper, through either a Foundry resource or an Azure OpenAI resource. Content Understanding isn't available from a standalone Azure OpenAI resource.
Also called Azure OpenAI in Foundry Models.
Read more: Microsoft Learn
In the Ultra Transcenders books
Each book explains Azure OpenAI in context, with comparison tables and the common traps.
Terms in this definition
- GPT
A family of large language models from OpenAI, based on the transformer architecture and pretrained on big text collections, used to understand and produce language and code.
- o-series
Reasoning models from OpenAI, namely o1, o3 and o4-mini, that think longer to tackle science, coding and maths problems. Azure has deprecated all of them, with retirement on 19 November 2026 and GPT-5.6 models as successors.
- Embeddings
Vectors of numbers encoding what text means, placing related content close together as judged by cosine similarity. Switching model or vector size requires embedding all content again.
- Whisper
An OpenAI model that transcribes speech and can translate it into English, with files capped at 25 MB; it is offered through Azure Speech and Azure OpenAI. Azure OpenAI's whisper (001) model is due to retire on 15 December 2026.
- Microsoft Foundry resource
Azure resource at the top of the Foundry hierarchy (
Microsoft.CognitiveServices/accounts, kindAIServices). It governs model deployments, networking, security, policy and shared connections, while child projects keep teams isolated. - Azure Content Understanding in Foundry Tools
Foundry Tool whose generative AI analyzers turn documents, images, audio and video into structured fields by extracting, classifying or generating values.
Related terms
- Abuse monitoring
Azure OpenAI mechanism that looks across a customer's traffic for recurring misuse and flags it for review by Microsoft. Individual responses are not filtered by it.
- AIServices (resource kind)
Setting kind to AIServices creates the multi-service Microsoft Foundry resource, which Microsoft recommends; Speech, Vision, Language, Azure OpenAI and Content Understanding are then reached with a single credential at a single endpoint.
- API key (Foundry)
Secret shared by callers of an Azure OpenAI or Microsoft Foundry endpoint, conferring full access without any role check. Passing it around eliminates separation of roles, and it offers no network isolation.
- API key (Foundry resource key)
Key generated alongside an Azure OpenAI or Foundry Tools resource and supplied through the api-key header. It is used only when keyless authentication with Microsoft Entra ID cannot be.
- Assistants API
Azure OpenAI's OpenAI-style API based on assistants, threads and runs. It was deprecated and retired on 26 August 2026, with Foundry agents using conversations and responses taking its place.
- Azure OpenAI AvailabilityRate
Metric for Azure OpenAI calculated as total calls less server errors, divided by total calls; it reveals unavailability on the service side, meaning HTTP 5xx responses.
- Azure OpenAI Embedding skill
Skill in an AI Search skillset that sends each chunk to an embedding model deployment (text-embedding-3-small, text-embedding-3-large or ada-002) during indexing to produce vectors; charges come from Azure OpenAI.
- Azure OpenAI On Your Data
Feature of Azure OpenAI, now deprecated, for grounding chat answers in your own content retrieved from a search index. Its replacement for new work is Foundry Agent Service paired with Azure AI Search or Foundry IQ.