GPT-5 in Azure AI Foundry: The future of AI apps and agents starts here

We’re thrilled to announce that OpenAI’s latest flagship model, GPT-5, is now available on Azure AI Foundry. This isn’t just another model release; it’s the most advanced large language model (LLM) to date, setting new standards across various benchmarks.

For business leaders leveraging AI, the focus has shifted beyond simple chat functionalities. The expectation is now: can your AI truly generate content, reason through problems, and achieve measurable outcomes—safely and efficiently at scale?

Introducing OpenAI’s flagship model, GPT-5, in Azure AI Foundry. This model stands apart as the most potent LLM in the field, combining superior reasoning capabilities with high-performance generation at a cost-effective level. Built on Microsoft Azure’s enterprise-grade infrastructure, it enables organizations to transition from pilot projects to full production seamlessly.

GPT-5 in Azure AI Foundry: Designed for Practical Applications

Within Azure AI Foundry, you can access GPT-5 models via API, managed through a model router. The GPT-5 series boasts unique strengths:

• GPT-5 is a comprehensive reasoning model that excels in analytics and complex tasks, such as code generation, supporting a 272k token context.
• GPT-5 mini offers real-time capabilities for applications and agents that require reasoning and tool utilization to assist customers.
• GPT-5 nano represents a new generation of reasoning models aimed at ultra-low latency and quick responses, with robust Q&A features.
• GPT-5 chat facilitates fluid, multimodal conversations that maintain context throughout agent workflows, supporting a 128k token context.

Together, these options offer a seamless transition from advanced coding tasks to straightforward Q&A, all accessible through the Azure AI Foundry endpoint using the model router.

Beneath the surface, GPT-5 merges advanced reasoning, code generation, and natural language interaction. It provides analytical depth paired with intuitive dialogue, allowing it to solve complex problems and explain its methodologies. Its capabilities support multi-step processes and make transparent, auditable decisions. As a top-tier coding tool, GPT-5 can devise intricate workflows, assist in code migrations and refactoring, plus produce tests and documentation with clear explanations. With developer controls like reasoning_effort and verbosity to fine-tune depth and detail, teams can enhance their workflow, while new flexible tool-calling features expand tool compatibility without strict schemas.

Orchestrate with the Model Router—Then Scale with Agents

The introduction of GPT-5 on Azure AI Foundry signifies a significant advancement for the platform. Developers can now leverage the model router in Foundry Models to fully exploit the capabilities of the GPT-5 family (and other models) while saving up to 60% on inference costs with maintained precision. Utilising a finely-tuned SLM, the model router evaluates each request and selects the best model based on the task’s complexity and performance needs. Let the model router choose the ideal model, streamlining the creation of your AI applications.

And orchestration doesn’t end with routing—Foundry incorporates the same intelligence into its agents. Soon, GPT-5 will be part of the Foundry Agent Service, pairing advanced models with built-in tools, including new browser automation and Model Context Protocol (MCP) integrations. This results in policy-driven agents capable of searching, acting within web applications, and completing comprehensive tasks—coupled with Foundry telemetry and aligning with Microsoft’s Responsible AI standards.

Driving Business Impact with GPT-5

These advancements translate directly into business benefits.

In areas like research and knowledge work, GPT-5 enhances financial and legal analysis, market intelligence gathering, and due diligence processes—handling vast amounts of data and generating actionable insights. In operational contexts, it boosts logistics, risk management, and claims processing by marrying strong reasoning with policy compliance. Copilots and customer service teams gain from real-time, multi-turn, multimodal agents that reason on the fly, call for tools, complete tasks, and provide relevant context when reverting to human teams.

In software development, GPT-5 shines at generating code, modernising applications, and ensuring quality—improving coding style and clarity, thus speeding up review cycles.

For scenarios where cost or response time is critical, GPT-5 nano’s ultra-low latency framework provides fast, accurate answers, making it the ideal choice for fine-tuning and the go-to model for high-volume, straightforward requests.

GPT-5 Customer Highlights

Businesses are leveraging GPT-5 for intricate, vital tasks—boosting decision-making, enhancing coding processes, and driving product innovation.

SAP

SAP is thrilled to be among the first to harness the power of GPT-5 within our generative AI hub in AI Foundation. This model will empower our product and developer teams to deliver impactful innovations to our customers.

—Dr. Walter Sun, SVP and Global Head of AI, SAP SE

Relativity

The GPT-5 in Azure AI Foundry sets a new standard for applying legal data intelligence… This next-gen AI will enable legal teams to extract deeper insights, speed up decision-making, and bolster strategies throughout the legal process.

—Dr. Aron Ahmadia, Senior Director, Applied Science, Relativity

Hebbia

The collaboration between Hebbia and Azure AI Foundry gives finance professionals an unparalleled advantage. With GPT-5’s advanced reasoning, they can identify critical data across numerous documents while swiftly conducting detailed financial analyses.

—Danny Wheller, VP of Business and Strategy

Developing with AI in GitHub Copilot and Visual Studio Code

GPT-5 is now rolling out to millions of developers using GitHub Copilot and Visual Studio Code, applying its advanced reasoning to tackle more complex challenges—from intricate refactoring to better navigating extensive codebases. GPT-5 helps developers code, test, and deploy faster, enhancing agent-driven coding tasks and improving code style and overall quality. With GPT-5, developers don’t just code faster—they code better.

With today’s VS Code update, developers gain an even more robust coding experience directly within the editor: GitHub Copilot’s coding assistant is now more adept at autonomously managing tasks in the background. Plus, the Copilot chat feature boosts productivity, supporting multiple tools per chat request, and allowing users to revert workspace changes to earlier states. We are also unveiling an enhanced extension for agent development using Azure AI Foundry directly within the VS Code environment.

These updates reflect Microsoft’s vision to revolutionise software development with AI, integrating advanced AI capabilities into the entire software lifecycle.

Built-in Security, Safety, and Governance

Security and safety are multi-layered protections essential in every domain, including AI. We consider these layers with the model at the centre, ensuring that GPT-5 is more secure than ever:

The Microsoft AI Red Team has found GPT-5 to possess one of the strongest safety profiles among OpenAI models, performing on par with, or better than, o3.

—Dr. Sarah Bird, Chief Product Officer of Responsible AI, Microsoft

In considering the layers around this core—Azure AI Foundry introduces several crucial controls:

• Azure AI Content Safety protections are integrated into every prompt and completion, including prompt shields that help detect and reduce prompt injection attempts before reaching the model.

• Lastly, security signals link directly with Microsoft Defender for Cloud, while runtime metadata and evaluation results integrate with Microsoft Purview for audits, data-loss prevention, and regulatory compliance, extending protection throughout the entire GPT-5 lifecycle.

Integrating AI into Every Workflow with GitHub Copilot and Visual Studio Code

Starting today, GPT-5 is rolling out to millions of developers using GitHub Copilot and Visual Studio Code, who can now choose GPT-5 to assist in writing, testing, and deploying code—and develop agents using the Azure AI Foundry extension, all within the VS Code environment. GPT-5 supports complex coding tasks with remarkable enhancements to coding style, usability, and quality, which are highly sought-after improvements within the developer community.

Our assessments suggest OpenAI’s GPT-5 reasoning capabilities and contextual awareness surpass those of o3, empowering developers to tackle more intricate challenges—ranging from refactoring to efficiently navigating extensive codebases. With GPT-5, users in the Visual Studio suite can not only code faster but also elevate their coding perfection.

VS Code, along with our recent decision to open-source GitHub Copilot, exemplifies our dedication to open tools and standards, demonstrating our ability to keep pace with rapid model innovations whilst prioritising the developer experience. In today’s release of VS Code, developers have even greater control over their interaction within chats, with improved reliability for terminal tools, updates to the tool picker and limits, along with checkpoint features.

Today’s announcement broadens Microsoft’s mission to transform software development using AI, integrating advanced AI capabilities throughout the software lifecycle.

Get Started Today

GPT-5 is now accessible via our Standard offering in Azure AI Foundry, featuring deployment choices tailored for cost-efficiency and governance, including Global and Data Zone (United States, European Union) options for data residency and compliance.¹

With Azure AI Foundry’s unparalleled reliability, real-time assessments, inherent observability, and secure deployment options, you can transition confidently from pilot to production—all enhanced by tools like Model Router to optimise quality, latency, and costs across workloads.

---

¹Pricing is accurate as of August 2025