{
    "meta": {
        "timestamp": 1761337961,
        "hash": "6450b5f76bfc",
        "file": "posttext.txt",
        "language": "auto",
        "schema_type": "Document"
    },
    "content": {
        "text": "Skip to main contentSkip to Ask Learn chat experience\r\nMicrosoft Ignite\r\nNovember 17–21, 2025\r\n\r\nLearn\r\nSign in\r\nLearn  Azure  AI Foundry \r\nFoundry Models sold directly by Azure\r\n09\/18\/2025\r\nChoose a collection from Foundry Models sold directly by Azure\r\nIn this article\r\nAzure OpenAI in Azure AI Foundry models\r\nGPT-5\r\ngpt-oss\r\nGPT-4.1 series\r\nShow 16 more\r\nThis article lists a selection of Azure AI Foundry Models sold directly by Azure along with their capabilities, deployment types, and regions of availability, excluding deprecated and legacy models. Models sold directly by Azure include all Azure OpenAI models and specific, selected models from top providers.\r\n\r\nDepending on the kind of project you use in Azure AI Foundry, you see a different selection of models. Specifically, if you use a Foundry project built on an Azure AI Foundry resource, you see the models that are available for standard deployment to a Foundry resource. Alternatively, if you use a hub-based project hosted by an Azure AI Foundry hub, you see models that are available for deployment to managed compute and serverless APIs. These model selections often overlap because many models support multiple deployment options.\r\n\r\nTo learn more about attributes of Foundry Models sold directly by Azure, see Explore Azure AI Foundry Models.\r\n\r\n Note\r\n\r\nFoundry Models sold directly by Azure also include select models from the following top model providers:\r\n\r\nBlack Forest Labs: FLUX.1-Kontext-pro, FLUX-1.1-pro\r\nDeepSeek: DeepSeek-V3.1, DeepSeek-V3-0324, DeepSeek-R1-0528, DeepSeek-R1\r\nMeta: Llama-4-Maverick-17B-128E-Instruct-FP8, Llama-3.3-70B-Instruct\r\nMicrosoft: MAI-DS-R1\r\nMistral: mistral-document-ai-2505\r\nxAI: grok-code-fast-1, grok-3, grok-3-mini, grok-4-fast-reasoning, grok-4-fast-non-reasoning, grok-4\r\nTo learn about these models, switch to Other model collections at the top of this article.\r\n\r\nAzure OpenAI in Azure AI Foundry models\r\nAzure OpenAI is powered by a diverse set of models with different capabilities and price points. Model availability varies by region and cloud. For Azure Government model availability, refer to Azure OpenAI in Azure Government.\r\n\r\nModels\tDescription\r\nSora\tNEW sora-2\r\nGPT-5 series\tNEW gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-chat\r\ngpt-oss\tNEW open-weight reasoning models\r\ncodex-mini\tFine-tuned version of o4-mini.\r\nGPT-4.1 series\tgpt-4.1, gpt-4.1-mini, gpt-4.1-nano\r\nmodel-router\tA model that intelligently selects from a set of underlying chat models to respond to a given prompt.\r\ncomputer-use-preview\tAn experimental model trained for use with the Responses API computer use tool.\r\no-series models\tReasoning models with advanced problem solving and increased focus and capability.\r\nGPT-4o, GPT-4o mini, and GPT-4 Turbo\tCapable Azure OpenAI models with multimodal versions, which can accept both text and images as input.\r\nGPT-4\tA set of models that improve on GPT-3.5 and can understand and generate natural language and code.\r\nGPT-3.5\tA set of models that improve on GPT-3 and can understand and generate natural language and code.\r\nEmbeddings\tA set of models that can convert text into numerical vector form to facilitate text similarity.\r\nImage generation\tA series of models that can generate original images from natural language.\r\nVideo generation\tA model that can generate original video scenes from text instructions.\r\nAudio\tA series of models for speech to text, translation, and text to speech. GPT-4o audio models support either low latency speech in, speech out conversational interactions or audio generation.\r\nGPT-5\r\nRegion availability\r\nModel\tRegion\r\ngpt-5 (2025-08-07)\tSee the models table.\r\ngpt-5-mini (2025-08-07)\tSee the models table.\r\ngpt-5-nano (2025-08-07)\tSee the models table.\r\ngpt-5-chat (2025-08-07)\tSee the models table.\r\ngpt-5-chat (2025-10-03)\tEast US2 (Global Standard) and Sweden Central (Global Standard)\r\ngpt-5-codex (2025-09-11)\tEast US2 (Global Standard) and Sweden Central (Global Standard)\r\ngpt-5-pro (2025-10-06)\tEast US2 (Global Standard) and Sweden Central (Global Standard)\r\nRegistration is required for access to the gpt-5-pro, gpt-5, & gpt-5-codex models.\r\n\r\ngpt-5-mini, gpt-5-nano, and gpt-5-chat do not require registration.\r\n\r\nAccess will be granted based on Microsoft's eligibility criteria. Customers who previously applied and received access to o3, don't need to reapply as their approved subscriptions will automatically be granted access upon model release.\r\n\r\nModel ID\tDescription\tContext Window\tMax Output Tokens\tTraining Data (up to)\r\ngpt-5 (2025-08-07)\t- Reasoning\r\n- Chat Completions API.\r\n- Responses API.\r\n- Structured outputs.\r\n- Text and image processing.\r\n- Functions, tools, and parallel tool calling.\r\n- Full summary of capabilities.\t400,000\r\n\r\nInput: 272,000\r\nOutput: 128,000\t128,000\tSeptember 30, 2024\r\ngpt-5-mini (2025-08-07)\t- Reasoning\r\n- Chat Completions API.\r\n- Responses API.\r\n- Structured outputs.\r\n- Text and image processing.\r\n- Functions, tools, and parallel tool calling.\r\n- Full summary of capabilities.\t400,000\r\n\r\nInput: 272,000\r\nOutput: 128,000\t128,000\tMay 31, 2024\r\ngpt-5-nano (2025-08-07)\t- Reasoning\r\n- Chat Completions API.\r\n- Responses API.\r\n- Structured outputs.\r\n- Text and image processing.\r\n- Functions, tools, and parallel tool calling.\r\n- Full summary of capabilities.\t400,000\r\n\r\nInput: 272,000\r\nOutput: 128,000\t128,000\tMay 31, 2024\r\ngpt-5-chat (2025-08-07)\r\nPreview\t- Chat Completions API.\r\n- Responses API.\r\n- Input: Text\/Image\r\n- Output: Text only\t128,000\t16,384\tSeptember 30, 2024\r\ngpt-5-chat (2025-10-03)\r\nPreview1\t- Chat Completions API.\r\n- Responses API.\r\n- Input: Text\/Image\r\n- Output: Text only\t128,000\t16,384\tSeptember 30, 2024\r\ngpt-5-codex (2025-09-11)\t- Responses API only.\r\n- Input: Text\/Image\r\n- Output: Text only\r\n- Structured outputs.\r\n- Text and image processing.\r\n- Functions, tools, and parallel tool calling.\r\n- Full summary of capabilities\r\n- Optimized for Codex CLI & Codex VS Code extension\t400,000\r\n\r\nInput: 272,000\r\nOutput: 128,000\t128,000\t-\r\ngpt-5-pro (2025-10-06)\t- Reasoning\r\n- Responses API.\r\n- Structured outputs.\r\n- Text and image processing.\r\n- Functions and tools\r\n- Full summary of capabilities.\t400,000\r\n\r\nInput: 272,000\r\nOutput: 128,000\t128,000\tSeptember 30, 2024\r\n Note\r\n\r\n1 gpt-5-chat version 2025-10-03 introduces a significant enhancement focused on emotional intelligence and mental health capabilities. This upgrade integrates specialized datasets and refined response strategies to improve the model’s ability to:\r\n\r\nUnderstand and interpret emotional context more accurately, enabling nuanced and empathetic interactions.\r\nProvide supportive, responsible responses in conversations related to mental health, ensuring sensitivity and adherence to best practices.\r\nThese improvements aim to make GPT-5-chat more context-aware, human-centric, and reliable in scenarios where emotional tone and well-being considerations are critical.\r\n\r\ngpt-oss\r\nRegion availability\r\nModel\tRegion\r\ngpt-oss-120b\tAll Azure OpenAI regions\r\nCapabilities\r\nModel ID\tDescription\tContext Window\tMax Output Tokens\tTraining Data (up to)\r\ngpt-oss-120b (Preview)\t- Text in\/text out only\r\n- Chat Completions API\r\n- Streaming\r\n- Function calling\r\n- Structured outputs\r\n- Reasoning\r\n- Available for deployment1 and via managed compute\t131,072\t131,072\tMay 31, 2024\r\ngpt-oss-20b (Preview)\t- Text in\/text out only\r\n- Chat Completions API\r\n- Streaming\r\n- Function calling\r\n- Structured outputs\r\n- Reasoning\r\n- Available via managed compute and Foundry Local\t131,072\t131,072\tMay 31, 2024\r\n1 Unlike other Azure OpenAI models gpt-oss-120b requires an Azure AI Foundry project to deploy the model.\r\n\r\nDeploy with code\r\ncli\r\n\r\nCopy\r\naz cognitiveservices account deployment create \\\r\n  --name \"Foundry-project-resource\" \\\r\n  --resource-group \"test-rg\" \\\r\n  --deployment-name \"gpt-oss-120b\" \\\r\n  --model-name \"gpt-oss-120b\" \\\r\n  --model-version \"1\" \\\r\n  --model-format \"OpenAI-OSS\" \\\r\n  --sku-capacity 10 \\\r\n  --sku-name \"GlobalStandard\"\r\nGPT-4.1 series\r\nRegion availability\r\nModel\tRegion\r\ngpt-4.1 (2025-04-14)\tSee the models table.\r\ngpt-4.1-nano (2025-04-14)\tSee the models table.\r\ngpt-4.1-mini (2025-04-14)\tSee the models table.\r\nCapabilities\r\n Important\r\n\r\nA known issue is affecting all GPT 4.1 series models. Large tool or function call definitions that exceed 300,000 tokens will result in failures, even though the 1 million token context limit of the models wasn't reached.\r\n\r\nThe errors can vary based on API call and underlying payload characteristics.\r\n\r\nHere are the error messages for the Chat Completions API:\r\n\r\nError code: 400 - {'error': {'message': \"This model's maximum context length is 300000 tokens. However, your messages resulted in 350564 tokens (100 in the messages, 350464 in the functions). Please reduce the length of the messages or functions.\", 'type': 'invalid_request_error', 'param': 'messages', 'code': 'context_length_exceeded'}}\r\n\r\nError code: 400 - {'error': {'message': \"Invalid 'tools[0].function.description': string too long. Expected a string with maximum length 1048576, but got a string with length 2778531 instead.\", 'type': 'invalid_request_error', 'param': 'tools[0].function.description', 'code': 'string_above_max_length'}}\r\n\r\nHere's the error message for the Responses API:\r\n\r\nError code: 500 - {'error': {'message': 'The server had an error processing your request. Sorry about that! You can retry your request, or contact us through an Azure support request at: https:\/\/go.microsoft.com\/fwlink\/?linkid=2213926 if you keep seeing this error. (Please include the request ID d2008353-291d-428f-adc1-defb5d9fb109 in your email.)', 'type': 'server_error', 'param': None, 'code': None}}\r\nModel ID\tDescription\tContext window\tMax output tokens\tTraining data (up to)\r\ngpt-4.1 (2025-04-14)\t- Text and image input\r\n- Text output\r\n- Chat completions API\r\n- Responses API\r\n- Streaming\r\n- Function calling\r\n- Structured outputs (chat completions)\t- 1,047,576\r\n- 128,000 (provisioned managed deployments)\r\n- 300,000 (batch deployments)\t32,768\tMay 31, 2024\r\ngpt-4.1-nano (2025-04-14)\t- Text and image input\r\n- Text output\r\n- Chat completions API\r\n- Responses API\r\n- Streaming\r\n- Function calling\r\n- Structured outputs (chat completions)\t- 1,047,576\r\n- 128,000 (provisioned managed deployments)\r\n- 300,000 (batch deployments)\t32,768\tMay 31, 2024\r\ngpt-4.1-mini (2025-04-14)\t- Text and image input\r\n- Text output\r\n- Chat completions API\r\n- Responses API\r\n- Streaming\r\n- Function calling\r\n- Structured outputs (chat completions)\t- 1,047,576\r\n- 128,000 (provisioned managed deployments)\r\n- 300,000 (batch deployments)\t32,768\tMay 31, 2024\r\nmodel-router\r\nA model that intelligently selects from a set of underlying chat models to respond to a given prompt.\r\n\r\nRegion availability\r\nModel\tRegion\r\nmodel-router (2025-08-07)\tEast US 2 (Global Standard & Data Zone Standard), Sweden Central (Global Standard & Data Zone Standard)\r\nmodel-router (2025-05-19)\tEast US 2 (Global Standard & Data Zone Standard), Sweden Central (Global Standard & Data Zone Standard)\r\nBilling for Data Zone Standard model router deployments will begin no earlier than November 1, 2025.\r\n\r\nCapabilities\r\nModel ID\tDescription\tContext window\tMax output tokens\tTraining data (up to)\r\nmodel-router (2025-08-07)\tA model that intelligently selects from a set of underlying models to respond to a given prompt.\t200,000\t32,768 (GPT-4.1 series)\r\n100,000 (o4-mini)\r\n128,000 (gpt-5 reasoning models)\r\n16,384 (gpt-5-chat)\t-\r\nmodel-router (2025-05-19)\tA model that intelligently selects from a set of underlying chat models to respond to a given prompt.\t200,000\t32,768 (GPT-4.1 series)\r\n100,000 (o4-mini)\tMay 31, 2024\r\nLarger context windows are compatible with some of the underlying models. That means an API call with a larger context succeeds only if the prompt happens to be routed to the right model. Otherwise, the call fails.\r\n\r\ncomputer-use-preview\r\nAn experimental model trained for use with the Responses API computer use tool.\r\n\r\nIt can be used with third-party libraries to allow the model to control mouse and keyboard input, while getting context from screenshots of the current environment.\r\n\r\n Caution\r\n\r\nWe don't recommend using preview models in production. We'll upgrade all deployments of preview models to either future preview versions or to the latest stable, generally available version. Models that are designated preview don't follow the standard Azure OpenAI model lifecycle.\r\n\r\nRegistration is required to access computer-use-preview. Access is granted based on Microsoft's eligibility criteria. Customers who have access to other limited access models still need to request access for this model.\r\n\r\nTo request access, go to computer-use-preview limited access model application. When access is granted, you need to create a deployment for the model.\r\n\r\nRegion availability\r\nModel\tRegion\r\ncomputer-use-preview\tSee the models table.\r\nCapabilities\r\nModel ID\tDescription\tContext window\tMax output tokens\tTraining data (up to)\r\ncomputer-use-preview (2025-03-11)\tSpecialized model for use with the Responses API computer use tool\r\n\r\n- Tools\r\n- Streaming\r\n- Text (input\/output)\r\n- Image (input)\t8,192\t1,024\tOctober 2023\r\no-series models\r\nThe Azure OpenAI o-series models are designed to tackle reasoning and problem-solving tasks with increased focus and capability. These models spend more time processing and understanding the user's request, making them exceptionally strong in areas like science, coding, and math, compared to previous iterations.\r\n\r\nModel ID\tDescription\tMax request (tokens)\tTraining data (up to)\r\ncodex-mini (2025-05-16)\tFine-tuned version of o4-mini.\r\n- Responses API.\r\n- Structured outputs.\r\n- Text and image processing.\r\n- Functions and tools.\r\nFull summary of capabilities.\tInput: 200,000\r\nOutput: 100,000\tMay 31, 2024\r\no3-pro (2025-06-10)\t- Responses API.\r\n- Structured outputs.\r\n- Text and image processing.\r\n- Functions and tools.\r\nFull summary of capabilities.\tInput: 200,000\r\nOutput: 100,000\tMay 31, 2024\r\no4-mini (2025-04-16)\t- New reasoning model, offering enhanced reasoning abilities.\r\n- Chat Completions API.\r\n- Responses API.\r\n- Structured outputs.\r\n- Text and image processing.\r\n- Functions and tools.\r\nFull summary of capabilities.\tInput: 200,000\r\nOutput: 100,000\tMay 31, 2024\r\no3 (2025-04-16)\t- New reasoning model, offering enhanced reasoning abilities.\r\n- Chat Completions API.\r\n- Responses API.\r\n- Structured outputs.\r\n- Text and image processing.\r\n- Functions, tools, and parallel tool calling.\r\nFull summary of capabilities.\tInput: 200,000\r\nOutput: 100,000\tMay 31, 2024\r\no3-mini (2025-01-31)\t- Enhanced reasoning abilities.\r\n- Structured outputs.\r\n- Text-only processing.\r\n- Functions and tools.\tInput: 200,000\r\nOutput: 100,000\tOctober 2023\r\no1 (2024-12-17)\t- Enhanced reasoning abilities.\r\n- Structured outputs.\r\n- Text and image processing.\r\n- Functions and tools.\tInput: 200,000\r\nOutput: 100,000\tOctober 2023\r\no1-preview (2024-09-12)\tOlder preview version.\tInput: 128,000\r\nOutput: 32,768\tOctober 2023\r\no1-mini (2024-09-12)\tA faster and more cost-efficient option in the o1 series, ideal for coding tasks that require speed and lower resource consumption.\r\n- Global Standard deployment available by default.\r\n- Standard (regional) deployments are currently only available for select customers who received access as part of the o1-preview limited access release.\tInput: 128,000\r\nOutput: 65,536\tOctober 2023\r\nTo learn more about advanced o-series models, see Getting started with reasoning models.\r\n\r\nRegion availability\r\nModel\tRegion\r\ncodex-mini\tEast US2 & Sweden Central (Global Standard).\r\no3-pro\tEast US2 & Sweden Central (Global Standard).\r\no4-mini\tSee the models table.\r\no3\tSee the models table.\r\no3-mini\tSee the models table.\r\no1\tSee the models table.\r\no1-preview\tSee the models table. This model is available only for customers who were granted access as part of the original limited access.\r\no1-mini\tSee the models table.\r\nGPT-4o and GPT-4 Turbo\r\nGPT-4o integrates text and images in a single model, which enables it to handle multiple data types simultaneously. This multimodal approach enhances accuracy and responsiveness in human-computer interactions. GPT-4o matches GPT-4 Turbo in English text and coding tasks while offering superior performance in non-English language tasks and vision tasks, setting new benchmarks for AI capabilities.\r\n\r\nHow do I access the GPT-4o and GPT-4o mini models?\r\nGPT-4o and GPT-4o mini are available for Standard and Global Standard model deployment.\r\n\r\nYou need to create or use an existing resource in a supported Standard or Global Standard region where the model is available.\r\n\r\nWhen your resource is created, you can deploy the GPT-4o models. If you're performing a programmatic deployment, the model names are:\r\n\r\ngpt-4o version 2024-11-20\r\ngpt-4o version 2024-08-06\r\ngpt-4o version 2024-05-13\r\ngpt-4o-mini version 2024-07-18\r\nGPT-4 Turbo\r\nGPT-4 Turbo is a large multimodal model (accepting text or image inputs and generating text) that can solve difficult problems with greater accuracy than any of OpenAI's previous models. Like GPT-3.5 Turbo, and older GPT-4 models, GPT-4 Turbo is optimized for chat and works well for traditional completions tasks.\r\n\r\nGPT-4\r\nGPT-4 is the predecessor to GPT-4 Turbo. Both the GPT-4 and GPT-4 Turbo models have a base model name of gpt-4. You can distinguish between the GPT-4 and Turbo models by examining the model version.\r\n\r\nGPT-4 and GPT-4 Turbo models\r\nThese models can be used only with the Chat Completions API.\r\n\r\nSee Model versions to learn about how Azure OpenAI handles model version upgrades. See Working with models to learn how to view and configure the model version settings of your GPT-4 deployments.\r\n\r\nModel ID\tDescription\tMax request (tokens)\tTraining data (up to)\r\ngpt-4o (2024-11-20)\r\nGPT-4o (Omni)\t- Structured outputs.\r\n- Text and image processing.\r\n- JSON Mode.\r\n- Parallel function calling.\r\n- Enhanced accuracy and responsiveness.\r\n- Parity with English text and coding tasks compared to GPT-4 Turbo with Vision.\r\n- Superior performance in non-English languages and in vision tasks.\r\n- Enhanced creative writing ability.\tInput: 128,000\r\nOutput: 16,384\tOctober 2023\r\ngpt-4o (2024-08-06)\r\nGPT-4o (Omni)\t- Structured outputs.\r\n- Text and image processing.\r\n- JSON Mode.\r\n- Parallel function calling.\r\n- Enhanced accuracy and responsiveness.\r\n- Parity with English text and coding tasks compared to GPT-4 Turbo with Vision.\r\n- Superior performance in non-English languages and in vision tasks.\tInput: 128,000\r\nOutput: 16,384\tOctober 2023\r\ngpt-4o-mini (2024-07-18)\r\nGPT-4o mini\t- Fast, inexpensive, capable model ideal for replacing GPT-3.5 Turbo series models.\r\n- Text and image processing.\r\n- JSON Mode.\r\n- Parallel function calling.\tInput: 128,000\r\nOutput: 16,384\tOctober 2023\r\ngpt-4o (2024-05-13)\r\nGPT-4o (Omni)\t- Text and image processing.\r\n- JSON Mode.\r\n- Parallel function calling.\r\n- Enhanced accuracy and responsiveness.\r\n- Parity with English text and coding tasks compared to GPT-4 Turbo with Vision.\r\n- Superior performance in non-English languages and in vision tasks.\tInput: 128,000\r\nOutput: 4,096\tOctober 2023\r\ngpt-4 (turbo-2024-04-09)\r\nGPT-4 Turbo with Vision\tNew generally available model.\r\n- Replacement for all previous GPT-4 preview models (vision-preview, 1106-Preview, 0125-Preview).\r\n- Feature availability is currently different, depending on the method of input and the deployment type.\tInput: 128,000\r\nOutput: 4,096\tDecember 2023\r\n Caution\r\n\r\nWe don't recommend that you use preview models in production. We'll upgrade all deployments of preview models to either future preview versions or to the latest stable, generally available version. Models that are designated preview don't follow the standard Azure OpenAI model lifecycle.\r\n\r\nGPT-3.5\r\nGPT-3.5 models can understand and generate natural language or code. The most capable and cost effective model in the GPT-3.5 family is GPT-3.5 Turbo, which is optimized for chat and also works well for traditional completions tasks. GPT-3.5 Turbo is available for use with the Chat Completions API. GPT-3.5 Turbo Instruct has similar capabilities to text-davinci-003 when you use the Completions API instead of the Chat Completions API. We recommend using GPT-3.5 Turbo and GPT-3.5 Turbo Instruct over legacy GPT-3.5 and GPT-3 models.\r\n\r\nModel ID\tDescription\tMax request (tokens)\tTraining data (up to)\r\ngpt-35-turbo (0125) new\t- JSON Mode.\r\n- Parallel function calling.\r\n- Reproducible output (preview).\r\n- Higher accuracy when it responds in requested formats.\r\n- Includes a fix for a bug that caused a text-encoding issue for non-English language function calls.\tInput: 16,385\r\nOutput: 4,096\tSep 2021\r\ngpt-35-turbo (1106)\tOlder generally available model.\r\n- JSON Mode.\r\n- Parallel function calling.\r\n- Reproducible output (preview).\tInput: 16,385\r\nOutput: 4,096\tSep 2021\r\ngpt-35-turbo-instruct (0914)\tCompletions endpoint only.\r\n- Replacement for legacy completions models.\t4,097\tSep 2021\r\nTo learn more about how to interact with GPT-3.5 Turbo and the Chat Completions API, check out our in-depth how-to article.\r\n\r\nEmbeddings\r\ntext-embedding-3-large is the latest and most capable embedding model. You can't upgrade between embeddings models. To move from using text-embedding-ada-002 to text-embedding-3-large, you need to generate new embeddings.\r\n\r\ntext-embedding-3-large\r\ntext-embedding-3-small\r\ntext-embedding-ada-002\r\nOpenAI reports that testing shows that both the large and small third generation embeddings models offer better average multi-language retrieval performance with the MIRACL benchmark. They still maintain performance for English tasks with the MTEB benchmark.\r\n\r\nEvaluation benchmark\ttext-embedding-ada-002\ttext-embedding-3-small\ttext-embedding-3-large\r\nMIRACL average\t31.4\t44.0\t54.9\r\nMTEB average\t61.0\t62.3\t64.6\r\nThe third generation embeddings models support reducing the size of the embedding via a new dimensions parameter. Typically, larger embeddings are more expensive from a compute, memory, and storage perspective. When you can adjust the number of dimensions, you gain more control over overall cost and performance. The dimensions parameter isn't supported in all versions of the OpenAI 1.x Python library. To take advantage of this parameter, we recommend that you upgrade to the latest version: pip install openai --upgrade.\r\n\r\nOpenAI's MTEB benchmark testing found that even when the third generation model's dimensions are reduced to less than the 1,536 dimensions of text-embeddings-ada-002, performance remains slightly better.\r\n\r\nImage generation models\r\nThe image generation models generate images from text prompts that the user provides. GPT-image-1 series models are in limited access preview. DALL-E 3 is generally available for use with the REST APIs. DALL-E 2 and DALL-E 3 with client SDKs are in preview.\r\n\r\nRegistration is required to access gpt-image-1 or gpt-image-1-mini. Access is granted based on Microsoft's eligibility criteria. Customers who have access to other limited access models still need to request access for this model.\r\n\r\nTo request access, go to gpt-image-1 limited access model application. When access is granted, you need to create a deployment for the model.\r\n\r\nRegion availability\r\nModel\tRegion\r\ndall-e-3\tEast US\r\nAustralia East\r\nSweden Central\r\ngpt-image-1\tWest US 3 (Global Standard)\r\nEast US 2 (Global Standard)\r\nUAE North (Global Standard)\r\nPoland Central (Global Standard)\r\ngpt-image-1-mini\tEastUS (Global Standard)\r\nNorthCentralUS (Global Standard)\r\nVideo generation models\r\nSora is an AI model from OpenAI that can create realistic and imaginative video scenes from text instructions. Sora is in preview.\r\n\r\nRegion availability\r\nModel\tRegion\r\nsora\tEast US 2 (Global Standard)\r\nSweden Central (Global Standard)\r\nsora-2\tEast US 2 (Global Standard)\r\nSweden Central (Global Standard)\r\nAudio models\r\nAudio models in Azure OpenAI are available via the realtime, completions, and audio APIs.\r\n\r\nGPT-4o audio models\r\nThe GPT-4o audio models are part of the GPT-4o model family and support either low-latency, speech in, speech out conversational interactions or audio generation.\r\n\r\n Caution\r\n\r\nWe don't recommend using preview models in production. We'll upgrade all deployments of preview models to either future preview versions or to the latest stable, generally available version. Models that are designated preview don't follow the standard Azure OpenAI model lifecycle.\r\n\r\nDetails about maximum request tokens and training data are available in the following table:\r\n\r\nModel ID\tDescription\tMax request (tokens)\tTraining data (up to)\r\ngpt-4o-mini-audio-preview (2024-12-17)\r\nGPT-4o audio\tAudio model for audio and text generation.\tInput: 128,000\r\nOutput: 16,384\tSeptember 2023\r\ngpt-4o-audio-preview (2024-12-17)\r\nGPT-4o audio\tAudio model for audio and text generation.\tInput: 128,000\r\nOutput: 16,384\tSeptember 2023\r\ngpt-4o-realtime-preview (2025-06-03)\r\nGPT-4o audio\tAudio model for real-time audio processing.\tInput: 128,000\r\nOutput: 4,096\tOctober 2023\r\ngpt-4o-realtime-preview (2024-12-17)\r\nGPT-4o audio\tAudio model for real-time audio processing.\tInput: 128,000\r\nOutput: 4,096\tOctober 2023\r\ngpt-4o-mini-realtime-preview (2024-12-17)\r\nGPT-4o audio\tAudio model for real-time audio processing.\tInput: 128,000\r\nOutput: 4,096\tOctober 2023\r\ngpt-realtime (2025-08-28) (GA)\r\ngpt-realtime-mini (2025-10-06)\r\ngpt-audio(2025-08-28)\r\ngpt-audio-mini(2025-10-06)\tAudio model for real-time audio processing.\tInput: 28,672\r\nOutput: 4,096\tOctober 2023\r\nTo compare the availability of GPT-4o audio models across all regions, refer to the models table.\r\n\r\nAudio API\r\nThe audio models via the \/audio API can be used for speech to text, translation, and text to speech.\r\n\r\nSpeech-to-text models\r\nModel ID\tDescription\tMax request (audio file size)\r\nwhisper\tGeneral-purpose speech recognition model.\t25 MB\r\ngpt-4o-transcribe\tSpeech-to-text model powered by GPT-4o.\t25 MB\r\ngpt-4o-mini-transcribe\tSpeech-to-text model powered by GPT-4o mini.\t25 MB\r\ngpt-4o-transcribe-diarize\tSpeech-to-text model with automatic speech recognition.\t25 MB\r\nSpeech translation models\r\nModel ID\tDescription\tMax request (audio file size)\r\nwhisper\tGeneral-purpose speech recognition model.\t25 MB\r\nText-to-speech models (preview)\r\nModel ID\tDescription\r\ntts\tText-to-speech model optimized for speed.\r\ntts-hd\tText-to-speech model optimized for quality.\r\ngpt-4o-mini-tts\tText-to-speech model powered by GPT-4o mini.\r\n\r\nYou can guide the voice to speak in a specific style or tone.\r\nFor more information, see Audio models region availability later in this article.\r\n\r\nModel summary table and region availability\r\nModels by deployment type\r\nAzure OpenAI provides customers with choices on the hosting structure that fits their business and usage patterns. The service offers two main types of deployment:\r\n\r\nStandard: Has a global deployment option, routing traffic globally to provide higher throughput.\r\nProvisioned: Also has a global deployment option, allowing customers to purchase and deploy provisioned throughput units across Azure global infrastructure.\r\nAll deployments can perform the exact same inference operations, but the billing, scale, and performance are substantially different. To learn more about Azure OpenAI deployment types, see our Deployment types guide.\r\n\r\nGlobal Standard\r\nGlobal Provisioned managed\r\nGlobal Batch\r\nData Zone Standard\r\nData Zone Provisioned managed\r\nData Zone Batch\r\nStandard\r\nProvisioned managed\r\nGlobal Standard model availability\r\nRegion\tgpt-5, 2025-08-07\tgpt-5-mini, 2025-08-07\tgpt-5-nano, 2025-08-07\tgpt-5-chat, 2025-08-07\to3-pro, 2025-06-10\tcodex-mini, 2025-05-16\tsora, 2025-05-02\tmodel-router, 2025-08-07\tmodel-router, 2025-05-19\to3, 2025-04-16\to4-mini, 2025-04-16\tgpt-image-1, 2025-04-15\tgpt-image-1-mini, 2025-10-06\tgpt-4.1, 2025-04-14\tgpt-4.1-nano, 2025-04-14\tgpt-4.1-mini, 2025-04-14\tcomputer-use-preview, 2025-03-11\to3-mini, 2025-01-31\to1, 2024-12-17\to1-mini, 2024-09-12\tgpt-4o, 2024-05-13\tgpt-4o, 2024-08-06\tgpt-4o, 2024-11-20\tgpt-4o-mini, 2024-07-18\tgpt-4, turbo-2024-04-09\ttext-embedding-3-small, 1\ttext-embedding-3-large, 1\ttext-embedding-ada-002, 2\tgpt-4o-realtime-preview, 2024-12-17\tgpt-4o-realtime-preview, 2025-06-03\tgpt-4o-audio-preview, 2024-12-17\tgpt-4o-mini-realtime-preview, 2024-12-17\tgpt-4o-mini-audio-preview, 2024-12-17\tgpt-4o-transcribe, 2025-03-20\tgpt-4o-mini-tts, 2025-03-20\tgpt-4o-mini-transcribe, 2025-03-20\tgpt-5-codex, 2025-09-15\tgpt-audio, 2025-08-28\tgpt-realtime, 2025-08-28\to3-deep-research, 2025-06-26\r\naustraliaeast\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nbrazilsouth\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\ncanadaeast\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\neastus\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t✅\t-\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t✅\t-\t-\t-\t-\t-\t-\t-\t\r\neastus2\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t\r\nfrancecentral\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\ngermanywestcentral\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nitalynorth\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\njapaneast\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nkoreacentral\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nnorthcentralus\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t✅\t✅\t-\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nnorwayeast\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t\r\npolandcentral\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t✅\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nsouthafricanorth\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nsouthcentralus\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nsouthindia\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nspaincentral\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nswedencentral\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t✅\t-\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t✅\t-\t✅\t✅\t✅\t✅\t-\t\r\nswitzerlandnorth\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nuaenorth\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t✅\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nuksouth\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nwesteurope\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\nwestus\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t-\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t\r\nwestus3\t-\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t✅\t✅\t-\t✅\t-\t-\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t\r\n Note\r\n\r\no3-deep-research is currently only available with Azure AI Foundry Agent Service. To learn more, see the Deep Research tool guidance.\r\n\r\nThis table doesn't include fine-tuning regional availability information. Consult the fine-tuning section for this information.\r\n\r\nStandard deployment (regional) models by endpoint\r\nChat completions\r\nEmbeddings\r\nImage generation\r\nVideo generation\r\nAudio\r\nCompletions (legacy)\r\nChat completions\r\nRegion\to1-preview, 2024-09-12\to1-mini, 2024-09-12\tgpt-4o, 2024-05-13\tgpt-4o, 2024-11-20\tgpt-4o, 2024-08-06\tgpt-4o-mini, 2024-07-18\tgpt-4, turbo-2024-04-09\tgpt-35-turbo, 1106\tgpt-35-turbo, 0125\r\naustraliaeast\t-\t-\t-\t✅\t-\t-\t-\t✅\t✅\r\ncanadaeast\t-\t-\t-\t✅\t-\t-\t-\t✅\t✅\r\neastus\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t✅\r\neastus2\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t✅\r\nfrancecentral\t-\t-\t-\t✅\t-\t-\t-\t✅\t✅\r\njapaneast\t-\t-\t-\t✅\t-\t-\t-\t-\t✅\r\nnorthcentralus\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t✅\r\nnorwayeast\t-\t-\t-\t✅\t-\t-\t-\t-\t-\r\nsouthcentralus\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t✅\r\nsouthindia\t-\t-\t-\t✅\t-\t-\t-\t✅\t✅\r\nswedencentral\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\r\nswitzerlandnorth\t-\t-\t-\t✅\t-\t-\t-\t-\t✅\r\nuksouth\t-\t-\t-\t✅\t-\t-\t-\t✅\t✅\r\nwesteurope\t-\t-\t-\t-\t-\t-\t-\t-\t✅\r\nwestus\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t✅\r\nwestus3\t✅\t✅\t✅\t✅\t✅\t✅\t✅\t-\t✅\r\n Note\r\n\r\no1-mini is currently available to all customers for Global Standard deployment.\r\n\r\nSelect customers were granted standard (regional) deployment access to o1-mini as part of the o1-preview limited access release. At this time, access to o1-mini standard (regional) deployments isn't being expanded.\r\n\r\nGPT-4 and GPT-4 Turbo model availability\r\nGPT-3.5 models\r\nTo learn about how Azure OpenAI handles model version upgrades, see Model versions. To learn how to view and configure the model version settings of your GPT-3.5 Turbo deployments, see Working with models.\r\n\r\nFine-tuning models\r\n Note\r\n\r\ngpt-35-turbo: Fine-tuning of this model is limited to a subset of regions, and isn't available in every region the base model is available.\r\n\r\nThe supported regions for fine-tuning might vary if you use Azure OpenAI models in an Azure AI Foundry project versus outside a project.\r\n\r\nModel ID\tStandard training regions\tGlobal training\tMax request (tokens)\tTraining data (up to)\tModality\r\ngpt-35-turbo\r\n(1106)\tEast US2\r\nNorth Central US\r\nSweden Central\r\nSwitzerland West\t-\tInput: 16,385\r\nOutput: 4,096\tSep 2021\tText to text\r\ngpt-35-turbo\r\n(0125)\tEast US2\r\nNorth Central US\r\nSweden Central\r\nSwitzerland West\t-\t16,385\tSep 2021\tText to text\r\ngpt-4o-mini\r\n(2024-07-18)\tNorth Central US\r\nSweden Central\t✅\tInput: 128,000\r\nOutput: 16,384\r\nTraining example context length: 65,536\tOct 2023\tText to text\r\ngpt-4o\r\n(2024-08-06)\tEast US2\r\nNorth Central US\r\nSweden Central\t✅\tInput: 128,000\r\nOutput: 16,384\r\nTraining example context length: 65,536\tOct 2023\tText and vision to text\r\ngpt-4.1\r\n(2025-04-14)\tNorth Central US\r\nSweden Central\t✅\tInput: 128,000\r\nOutput: 16,384\r\nTraining example context length: 65,536\tMay 2024\tText and vision to text\r\ngpt-4.1-mini\r\n(2025-04-14)\tNorth Central US\r\nSweden Central\t✅\tInput: 128,000\r\nOutput: 16,384\r\nTraining example context length: 65,536\tMay 2024\tText to text\r\ngpt-4.1-nano (2025-04-14)\tNorth Central US\r\nSweden Central\t✅\tInput: 128,000\r\nOutput: 16,384\r\nTraining example context length: 32,768\tMay 2024\tText to text\r\no4-mini\r\n(2025-04-16)\tEast US2\r\nSweden Central\t-\tInput: 128,000\r\nOutput: 16,384\r\nTraining example context length: 65,536\tMay 2024\tText to text\r\n Note\r\n\r\nGlobal training provides more affordable training per token, but doesn't offer data residency. It's currently available to Azure OpenAI resources in the following regions:\r\n\r\nAustralia East\r\nBrazil South\r\nCanada Central\r\nCanada East\r\nEast US\r\nEast US2\r\nFrance Central\r\nGermany West Central\r\nItaly North\r\nJapan East (no vision support)\r\nKorea Central\r\nNorth Central US\r\nNorway East\r\nPoland Central (no 4.1-nano support)\r\nSoutheast Asia\r\nSouth Africa North\r\nSouth Central US\r\nSouth India\r\nSpain Central\r\nSweden Central\r\nSwitzerland West\r\nSwitzerland North\r\nUK South\r\nWest Europe\r\nWest US\r\nWest US3\r\nAssistants (preview)\r\nFor Assistants, you need a combination of a supported model and a supported region. Certain tools and capabilities require the latest models. The following models are available in the Assistants API, SDK, and Azure AI Foundry. The following table is for standard deployment. For information on provisioned throughput unit availability, see Provisioned throughput. The listed models and regions can be used with both Assistants v1 and v2. You can use Global Standard models if they're supported in the following regions.\r\n\r\nRegion\tgpt-4o, 2024-05-13\tgpt-4o, 2024-08-06\tgpt-4o-mini, 2024-07-18\tgpt-4, 0613\tgpt-4, 1106-Preview\tgpt-4, 0125-Preview\tgpt-4, turbo-2024-04-09\tgpt-4-32k, 0613\tgpt-35-turbo, 0613\tgpt-35-turbo, 1106\tgpt-35-turbo, 0125\tgpt-35-turbo-16k, 0613\r\naustraliaeast\t-\t-\t-\t✅\t✅\t-\t-\t✅\t✅\t✅\t✅\t✅\r\neastus\t✅\t✅\t✅\t-\t-\t✅\t✅\t-\t✅\t-\t✅\t✅\r\neastus2\t✅\t✅\t✅\t-\t✅\t-\t✅\t-\t✅\t-\t✅\t✅\r\nfrancecentral\t-\t-\t-\t✅\t✅\t-\t-\t✅\t✅\t✅\t-\t✅\r\njapaneast\t-\t-\t-\t-\t-\t-\t-\t-\t✅\t-\t✅\t✅\r\nnorwayeast\t-\t-\t-\t-\t✅\t-\t-\t-\t-\t-\t-\t-\r\nsouthindia\t-\t-\t-\t-\t✅\t-\t-\t-\t-\t✅\t✅\t-\r\nswedencentral\t✅\t✅\t✅\t✅\t✅\t-\t✅\t✅\t✅\t✅\t-\t✅\r\nuksouth\t-\t-\t-\t-\t✅\t✅\t-\t-\t✅\t✅\t✅\t✅\r\nwestus\t✅\t✅\t✅\t-\t✅\t-\t✅\t-\t-\t✅\t✅\t-\r\nwestus3\t✅\t✅\t✅\t-\t✅\t-\t✅\t-\t-\t-\t✅\t-\r\nModel retirement\r\nFor the latest information on model retirements, refer to the model retirement guide.\r\n\r\nRelated content\r\nFoundry Models from partners and community\r\nModel retirement and deprecation\r\nLearn more about working with Azure OpenAI models\r\nLearn more about Azure OpenAI\r\nLearn more about fine-tuning Azure OpenAI models\r\nFeedback\r\nWas this page helpful?\r\n\r\nAdditional resources\r\nDocumentation\r\n\r\nAzure OpenAI in Azure AI Foundry Models Quotas and Limits\r\n\r\nThis article features detailed descriptions and best practices on the quotas and limits for Azure OpenAI.\r\n\r\nAzure OpenAI reasoning models - GPT-5 series, o3-mini, o1, o1-mini - Azure OpenAI\r\n\r\nLearn how to use Azure OpenAI's advanced GPT-5 series, o3-mini, o1, & o1-mini reasoning models\r\n\r\nWhat is Azure OpenAI in Azure AI Foundry Models?\r\n\r\nApply advanced language models to variety of use cases with Azure OpenAI\r\n\r\nShow 4 more\r\nTraining\r\n\r\nModule\r\n\r\nCreate a large language model deployment - Training\r\n\r\nLean how to create a large language model deployment.\r\n\r\nEvents\r\n\r\nExplore Azure AI Foundry\r\n\r\nAug 20, 2 AM - Oct 31, 1 AM\r\n\r\nUnlock the power of Azure AI Foundry and take your generative AI skills to the next level.\r\n\r\nLevel up today\r\nManage cookies\r\nAI Disclaimer\r\nPrevious Versions\r\nBlog\r\nContribute\r\nPrivacy\r\nTerms of Use\r\nTrademarks\r\n© Microsoft 2025",
        "summary": "Skip to main contentSkip to Ask Learn chat experience\r\nMicrosoft Ignite\r\nNovember 17–21, 2025\r\n\r\nLearn\r\nSign in\r\nLearn  Azure  AI Foundry \r\nFoundry Models sold directly by Azure\r\n09\/18\/2025\r\nChoose a collection from Foundry Models sold directly by Azure\r\nIn this article\r\nAzure OpenAI in Azure AI Foundry models\r\nGPT-5\r\ngpt-oss\r\nGPT-4.1 series\r\nShow 16 more\r\nThis article lists a selection of Azure AI Foundry Models sold directly by Azure along with their capabilities, deployment types, and regions of availability, excluding deprecated and legacy models. Models sold directly by Azure include all Azure OpenAI models and specific, selected models from top providers.\r\n\r\nDepending on the kind of project you use in Azure AI Foundry, you see a different selection of models. Specifically, if you use a Foundry project built on an Azure AI Foundry resource, you see the models that are available for standard deployment to a Foundry resource. Alternatively, if you use a hub-based project hosted by an Azure AI Foundry hub, you see models that are available for deployment to managed compute and serverless APIs. These model selections often overlap because many models support multiple deployment options.\r\n\r\nTo learn more about attributes of Foundry Models sold directly by Azure, see Explore Azure AI Foundry Models.\r\n\r\n Note\r\n\r\nFoundry Models sold directly by Azure also include select models from the following top model providers:\r\n\r\nBlack Forest Labs: FLUX.1-Kontext-pro, FLUX-1.1-pro\r\nDeepS...",
        "keywords": [],
        "entities": []
    },
    "structure": {
        "@context": "https:\/\/schema.org",
        "@type": "Document",
        "headline": "Untitled",
        "description": "Raw content"
    },
    "ai_meta": {
        "token_est": 9547,
        "chars": 38188,
        "crawler_hint": "Basic raw text",
        "richness_score": 1,
        "embedding_ready": false
    },
    "original_document": "https:\/\/data.aisenseapi.com\/\/content\/2025\/10\/24\/content_1761337961_6450b5f76bfc_posttext.txt",
    "reasoning": "Fallback: JSON parse error after cleaning raw response. Check log for original LLM output."
}