{"id":546,"date":"2025-05-19T08:59:14","date_gmt":"2025-05-19T15:59:14","guid":{"rendered":"https:\/\/devblogs.microsoft.com\/foundry\/?p=546"},"modified":"2025-05-27T13:26:53","modified_gmt":"2025-05-27T20:26:53","slug":"announcing-developer-essentials-for-agents-and-apps-in-azure-ai-foundry","status":"publish","type":"post","link":"https:\/\/devblogs.microsoft.com\/foundry\/announcing-developer-essentials-for-agents-and-apps-in-azure-ai-foundry\/","title":{"rendered":"Announcing Developer Essentials for Agents and Apps in Azure AI Foundry"},"content":{"rendered":"<p>In today\u2019s fast-moving tech landscape, developers need clear, efficient tooling to keep pace. <a href=\"https:\/\/www.forrester.com\/blogs\/predictions-2024-artificial-intelligence\/\">Forrester<\/a> tells us that 85 percent of enterprises now run multi-model AI strategies, so rapid testing and deployment are essential. Yet with seemingly endless models, platforms and providers to choose from, teams still lose precious cycles on guesswork\u2014tuning prompts, wiring orchestration, and chasing performance issues. <a href=\"https:\/\/azure.microsoft.com\/products\/ai-foundry\/\">Azure AI Foundry<\/a> removes that friction. The latest release streamlines model selection, customization, and monitoring and brings the <a href=\"https:\/\/aka.ms\/AgentService_ACOM\">Foundry Agent Service<\/a> to general availability, so you can customize, deploy, and run single or multiagent solutions at scale from one familiar platform.<\/p>\n<h2>Azure AI Foundry at a glance<\/h2>\n<p><a href=\"https:\/\/ai.azure.com\/\">Azure AI Foundry<\/a> is Microsoft\u2019s secure, flexible platform for designing, customizing, and managing AI apps and agents. Everything\u2014models, agents, tools, and observability\u2014lives behind a single portal, SDK, and REST endpoint, so you can ship to cloud or edge with governance and cost controls in place from day one.<\/p>\n<p><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/Azure-AI-Foundry.png\"><img decoding=\"async\" class=\"wp-image-674 size-full\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/Azure-AI-Foundry.png\" alt=\"Azure AI Foundry Marketecture\" width=\"1753\" height=\"985\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/Azure-AI-Foundry.png 1753w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/Azure-AI-Foundry-300x169.png 300w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/Azure-AI-Foundry-1024x575.png 1024w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/Azure-AI-Foundry-768x432.png 768w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/Azure-AI-Foundry-1536x863.png 1536w\" sizes=\"(max-width: 1753px) 100vw, 1753px\" \/><\/a><\/p>\n<p><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/Foundry_stack_hero.png\"><iframe title=\"YouTube video player\" src=\"\/\/www.youtube.com\/embed\/GD7MnIwAxYM?si=HE0p7XgOnFyXIIak\" width=\"560\" height=\"315\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"><\/iframe><\/a><\/p>\n<h2>Why the new release matters<\/h2>\n<p>With this update, Foundry truly becomes a one-stop-shop. In addition to giving you access to more than 11,000 <a href=\"https:\/\/azure.microsoft.com\/products\/ai-model-catalog\">Foundry Models<\/a>, we are now expanding the models hosted and sold by Microsoft including select models from Meta, Mistral AI, DeepSeek, and Black Forest Labs &#8211; coming soon in preview and models from xAI available today. Reserved capacity and provisioned throughput give every model the same predictable performance and billing experience. <strong>Foundry Agent Service<\/strong> is now GA. You can host a single agent or orchestrate a group of agents, expose them through the Agent to Agent (A2A) protocol, and connect them to other services using Model Context Protocol (MCP) or OpenAPI. Quick-start templates, VS Code integration, and GitHub workflows shorten the journey from idea to production to minutes instead of weeks. All of this arrives wrapped in enterprise grade trust: RBAC, customer managed keys, network isolation, and a responsible AI toolchain are built in rather than bolted on.<\/p>\n<h2>What\u2019s new in detail<\/h2>\n<h3>Foundry Models<\/h3>\n<p><a href=\"https:\/\/aka.ms\/AzureAIFoundryModels\">Select model families<\/a> hosted and sold directly from Microsoft &#8211; Llama 3, Llama 4, Grok 3, Mistral OCR, Codestral, and Flux now sit alongside <a href=\"https:\/\/azure.microsoft.com\/products\/ai-services\/openai-service\/\">Azure OpenAI<\/a> in <a href=\"https:\/\/aka.ms\/AzureAIFoundryModels\">Foundry Models<\/a>. These models carry the SLAs, security, and compliance Azure customers expect from any Microsoft product. Starting next month, you will also be able to use reserved capacity across these models with Foundry Provisioned Throughput (PTU)\u2014including Azure OpenAI models and the new models hosted and sold by Microsoft. This shared capacity model makes it easier than ever to operate multiple models under a unified, predictable performance and billing framework. We are also expanding our partnership with Hugging Face to provide access to a wide range of frontier and open-source Foundry Models. Soon, expect over <strong>11,000 models<\/strong> and hundreds of pre-built agentic tools. To make model choice easier, a new <strong>model leaderboard<\/strong> ranks models by quality, cost, and throughput, and a smart <strong>model router<\/strong> automatically picks the best model for each request based on your latency and budget constraints. In our tests comparing use of model router versus direct use of GPT-4.1 in Foundry Models, we saw up-to 60% cost savings with similar accuracy. <a href=\"https:\/\/aka.ms\/Build2025\/FoundryLocal\">Foundry Local<\/a> enables you to run a growing catalog of edge-optimized models and agents directly on Windows or macOS\u2014ideal for offline or privacy-sensitive workloads.<\/p>\n<p><div style=\"width: 1920px;\" class=\"wp-video\"><video class=\"wp-video-shortcode\" id=\"video-546-1\" width=\"1920\" height=\"1080\" loop autoplay preload=\"auto\" controls=\"controls\"><source type=\"video\/mp4\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/model_choice.mp4?_=1\" \/><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/model_choice.mp4\">https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/model_choice.mp4<\/a><\/video><\/div><\/p>\n<h3>Fine-tuning and distillation<\/h3>\n<p>Developers can now <a href=\"https:\/\/aka.ms\/Build25\/FoundryFT\">fine-tune <strong>GPT-4.1-nano<\/strong>, <strong>o4-mini<\/strong>, and <strong>Llama 4<\/strong><\/a> in Foundry Models with reinforcement fine-tuning available for advanced reasoning tasks. A new low-cost <em>Developer Tier<\/em> removes hosting fees during experimentation.<\/p>\n<p><div style=\"width: 1920px;\" class=\"wp-video\"><video class=\"wp-video-shortcode\" id=\"video-546-2\" width=\"1920\" height=\"1080\" loop autoplay preload=\"auto\" controls=\"controls\"><source type=\"video\/mp4\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/fine-tuning-and-distillation.mp4?_=2\" \/><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/fine-tuning-and-distillation.mp4\">https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/fine-tuning-and-distillation.mp4<\/a><\/video><\/div><\/p>\n<h3>Agent development<\/h3>\n<p><a href=\"https:\/\/aka.ms\/Build25\/AgentService_GA\"><strong>Foundry Agent Service<\/strong> is now GA<\/a>. You can host a single agent or orchestrate a group of agents, expose them through the Agent-to-Agent (A2A) protocol. The service simplifies agent development by integrating seamlessly with data sources like Microsoft Bing, Microsoft SharePoint, Azure AI Search, and Microsoft Fabric via knowledge tools while supporting task automation through action tools like Azure Logic Apps, Azure Functions, and custom tools using OpenAPI and Model Context Protocol (MCP). Behind the scenes, <strong>Semantic Kernel and AutoGen<\/strong> are merging into one unified SDK, giving you a single, composable API for defining, chaining, and deploying agents locally or in the cloud with identical behavior.\nFoundry Agent Service also connects with a <a href=\"https:\/\/aka.ms\/Build25\/Multi-Agent_Workflows\">centralized catalog<\/a> of agent code samples that developers can easily customize. These include predefined instructions, actions, APIs, knowledge, and tools, allowing developers to quickly create and deploy agents with Azure AI Foundry.\nMeanwhile, <a href=\"https:\/\/azure.microsoft.com\/products\/ai-services\/ai-search\/\">Azure AI Search<\/a> has introduced <em><a href=\"https:\/\/aka.ms\/AgentRAG\">agentic retrieval<\/a><\/em>\u2014now in public preview\u2014to handle automated multiturn planning, retrieval, and synthesis. And if you prefer to talk rather than type, the new <a href=\"https:\/\/aka.ms\/VoiceLiveBuild2025-blog\"><strong>Voice Live API<\/strong> <\/a>brings real-time speech input and output in more than 150 locales, all through the same endpoint.<\/p>\n<p><div style=\"width: 1920px;\" class=\"wp-video\"><video class=\"wp-video-shortcode\" id=\"video-546-3\" width=\"1920\" height=\"1080\" loop autoplay preload=\"auto\" controls=\"controls\"><source type=\"video\/mp4\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/foundry-agent-service.mp4?_=3\" \/><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/foundry-agent-service.mp4\">https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/foundry-agent-service.mp4<\/a><\/video><\/div><\/p>\n<h3>Content &amp; media<\/h3>\n<p>A new <strong>video playground<\/strong>\u2014coming soon\u2014lets you <a href=\"http:\/\/aka.ms\/SoraBuildBlogFinal\">try Sora in Foundry Models<\/a>\u2014coming soon\u2014and other cutting-edge generation models without provisioning any resources. Once you have a prompt you like, you can export ready-to-run code snippets straight into VS Code. For document-heavy workflows, the <a href=\"https:\/\/aka.ms\/AAIContentUnderstanding-Build2025\"><em>Pro<\/em>-mode in <strong>Azure AI Content Understanding<\/strong><\/a> collapses what used to be multiple extraction calls and human validation into one streamlined operation\u2014for example, comparing insurance claims against contract terms in a single step.<\/p>\n<p><div style=\"width: 1920px;\" class=\"wp-video\"><video class=\"wp-video-shortcode\" id=\"video-546-4\" width=\"1920\" height=\"1080\" loop autoplay preload=\"auto\" controls=\"controls\"><source type=\"video\/mp4\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/content-understanding.mp4?_=4\" \/><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/content-understanding.mp4\">https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/content-understanding.mp4<\/a><\/video><\/div><\/p>\n<h3>Developer productivity<\/h3>\n<p>Onboarding now takes seconds and produces a proof-of-concept in minutes. The GA version of the <strong><a href=\"https:\/\/aka.ms\/Build25\/AzureAIFoundrySDKBlog\">Foundry REST API<\/a><\/strong> unifies model inference, agent operations, and evaluations behind one endpoint, with SDKs for Python, C#, JavaScript, and Java. The updated <strong><a href=\"https:\/\/marketplace.visualstudio.com\/items?itemName=TeamsDevApp.vscode-ai-foundry\">VS Code extension<\/a><\/strong> offers YAML IntelliSense, full agent CRUD, and a new \u201cOpen in VS Code\u201d button in the portal that preloads keys and sample code. We have also published <strong><a href=\"https:\/\/aka.ms\/Build25\/AzureAIFoundryTemplatesBlog\">AI templates<\/a><\/strong> for common scenarios and released updates to the <a href=\"https:\/\/aka.ms\/M365AgentsToolkitBlog\">Microsoft 365 Agents Toolkit<\/a> so you can publish agents created with Foundry to Microsoft Copilot, Teams, and beyond.<\/p>\n<p><div style=\"width: 1920px;\" class=\"wp-video\"><video class=\"wp-video-shortcode\" id=\"video-546-5\" width=\"1920\" height=\"1080\" loop autoplay preload=\"auto\" controls=\"controls\"><source type=\"video\/mp4\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/vs_code_extension.mp4?_=5\" \/><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/vs_code_extension.mp4\">https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/vs_code_extension.mp4<\/a><\/video><\/div><\/p>\n<h3>Deployment &amp; observability<\/h3>\n<p>Dynamic capacity allocation scales on demand, and you can supercharge massive AI workloads with the dynamic token quota via Batch Global and DataZone deployments with new enhancements allowing you to queue up to 1 trillion tokens. Built-in dashboards in <a href=\"https:\/\/aka.ms\/Foundry-Observability\">Foundry Observability <\/a> cover quality, cost, safety, and ROI from the first playground test to production, and they plug into GitHub Actions and Azure DevOps so evaluations run automatically on every commit. In addition, Foundry Agent Service includes robust <a href=\"https:\/\/aka.ms\/Build25\/AgentOps\">AgentOps<\/a> capabilities\u2014such as tracing, evaluation, and monitoring\u2014helping developers validate, observe, and optimize agent behavior with confidence.<\/p>\n<p><div style=\"width: 1920px;\" class=\"wp-video\"><video class=\"wp-video-shortcode\" id=\"video-546-6\" width=\"1920\" height=\"1080\" loop autoplay preload=\"auto\" controls=\"controls\"><source type=\"video\/mp4\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/AgentEvals.mp4?_=6\" \/><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/AgentEvals.mp4\">https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/05\/AgentEvals.mp4<\/a><\/video><\/div><\/p>\n<h2>Customers innovating with Azure AI Foundry<\/h2>\n<p>More than 70,000 enterprises and digital natives have adopted Azure AI Foundry. A couple examples:<\/p>\n<ul>\n<li><a href=\"https:\/\/www.microsoft.com\/customers\/story\/23999-heineken-azure\"><strong>Heineken<\/strong> <\/a>used Azure AI Foundry to create a multi-agent platform to help employees access data and information across the company in their native language. Tasks that used to take 10 to 15 minutes now take 5 to 10 seconds.<\/li>\n<li><a href=\"https:\/\/www.microsoft.com\/customers\/story\/23953-accenture-azure-ai-foundry\"><strong>Accenture<\/strong> <\/a>used Azure AI Foundry to develop a centralized solution for secure generative AI development, including Azure AI Search, Azure AI Content Safety, and <a href=\"https:\/\/azure.microsoft.com\/en-us\/products\/machine-learning\">Azure Machine Learning<\/a>. Azure Machine Learning was used for custom model training, and Azure AI Foundry Models were used for fine-tuning. Accenture has also deployed more than 75 generative AI use cases across clients, with over 16 solutions in full production.<\/li>\n<\/ul>\n<h2>Get started today<\/h2>\n<p>The easiest way to explore is through the <a href=\"https:\/\/ai.azure.com\">Azure AI Foundry portal<\/a>. From there you can <a href=\"https:\/\/learn.microsoft.com\/azure\/ai-foundry\/how-to\/develop\/sdk-overview\">install the SDK<\/a>, follow the <a href=\"https:\/\/learn.microsoft.com\/azure\/ai-foundry\/\">documentation<\/a> and <a href=\"https:\/\/learn.microsoft.com\/plans\/34mi6tezkd7em\">Microsoft Learn courses<\/a>, and add the <a href=\"https:\/\/marketplace.visualstudio.com\/items?itemName=TeamsDevApp.vscode-ai-foundry\">VS Code extension<\/a>\u2014all in a few clicks. Sample repositories walk you through everything from a simple chat bot to a fully managed agent fleet.<\/p>\n<p>If you\u2019re attending <em>Microsoft Build 2025<\/em>, drop by one of my sessions\u2014<strong>\u201c<a href=\"https:\/\/build.microsoft.com\/sessions\/BRK155?source=sessions\">Azure AI Foundry: The Agent Factory,<\/a>\u201d<\/strong>\u00a0<strong>\u201c<a href=\"https:\/\/build.microsoft.com\/sessions\/BRK154?source=sessions\">Developer Essentials for Agents and Apps in Azure AI Foundry<\/a>,\u201d<\/strong>\u00a0and <strong>\u201c<a href=\"https:\/\/build.microsoft.com\/sessions\/BRK149?source=sessions\">Foundry Agent Service: Transforming Workflows with Azure AI Foundry<\/a>.\u201d<\/strong><\/p>\n<p>Happy coding!<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Learn about the latest release of Azure AI Foundry, including new models, agent service GA, unified developer tools, and productivity enhancements for building, deploying, and managing AI agents and apps at scale.<\/p>\n","protected":false},"author":69224,"featured_media":560,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1,27],"tags":[32,31,33,34],"class_list":["post-546","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-microsoft-foundry","category-whats-new","tag-ai","tag-azure","tag-foundry","tag-microsoft-build"],"acf":[],"blog_post_summary":"<p>Learn about the latest release of Azure AI Foundry, including new models, agent service GA, unified developer tools, and productivity enhancements for building, deploying, and managing AI agents and apps at scale.<\/p>\n","_links":{"self":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts\/546","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/users\/69224"}],"replies":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/comments?post=546"}],"version-history":[{"count":0,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts\/546\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/media\/560"}],"wp:attachment":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/media?parent=546"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/categories?post=546"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/tags?post=546"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}