August 8, 2025
5 min read
Steve Sweetman
Microsoft announces GPT-5 general availability in Azure AI Foundry, delivering powerful reasoning, coding, and agent orchestration for real-world AI workloads.
GPT-5 in Azure AI Foundry: Unlocking the Future of AI Apps and Agents
Microsoft has announced the general availability of OpenAI’s new flagship model, GPT-5, in Azure AI Foundry. This release signifies a major advancement, positioning GPT-5 as the most powerful large language model (LLM) to date across key benchmarks. It's engineered to address the escalating demands of AI applications that require robust generation, reasoning, and safe, scalable delivery of measurable outcomes.GPT-5 in Azure AI Foundry: Built for Real-World Workloads
GPT-5 is accessible via API within Azure AI Foundry and is managed by a sophisticated model router. This architecture provides a spectrum of models designed for diverse tasks:- GPT-5: A comprehensive reasoning model with deep analytical capabilities, ideal for complex tasks such as code generation, supporting an extensive 272k token context.
- GPT-5 mini: Tailored for real-time app and agent experiences that necessitate reasoning and tool calling to effectively resolve customer problems.
- GPT-5 nano: Engineered for ultra-low latency and speed, offering rich Q&A capabilities, perfect for use cases where cost and speed are critical.
- GPT-5 chat: Enables natural, multimodal, multi-turn conversations with a 128k token context, maintaining context awareness throughout agent workflows. Collectively, these models offer a seamless progression from intricate agentic coding to straightforward question-answering, all accessible through a unified Azure AI Foundry endpoint. GPT-5 integrates advanced reasoning, code generation, and natural language interaction, blending analytical depth with intuitive dialogue to tackle end-to-end challenges and articulate its reasoning. Its agentic capabilities support multi-step tool usage and extended action chains with transparent, auditable decision-making processes. Developers can fine-tune model behavior using parameters such as
- Research and knowledge work: Accelerates financial and legal analysis, market intelligence gathering, and due diligence processes through scalable reading and traceable, decision-ready outputs.
- Operations and decisioning: Enhances logistics support, risk assessment, and claims processing by combining robust reasoning with strict policy adherence.
- Copilots and customer experience: Powers multi-turn, multimodal agents that reason in real-time, utilize tools, resolve tasks, and escalate to human agents with comprehensive context.
- Software engineering: Excels in code generation, application modernization, and quality engineering, improving code style and explanations to streamline review cycles.
- Cost- and latency-sensitive scenarios: GPT-5 nano delivers rapid, high-accuracy responses, making it ideal for high-volume, straightforward requests and fine-tuning applications.
- Microsoft’s AI Red Team has identified GPT-5 as possessing one of the strongest safety profiles among OpenAI models, performing comparably to or better than previous iterations.
- Azure AI Content Safety implements protections, such as prompt shields, to detect and mitigate prompt-injection attempts.
- Integrated agent evaluators and the AI Red Team continuously conduct alignment, bias, and security testing, streaming real-time metrics into Azure Monitor and Application Insights.
- Security signals integrate with Microsoft Defender for Cloud, and runtime metadata feeds into Microsoft Purview for auditing, data-loss prevention, and regulatory compliance.
- AI Agents are Broken: Can GPT-5 Fix Them?
- Understanding AI Agent Washing Risks and Realities
- AI-Driven Crypto Trading Tools Reshape Market Strategies in 2025
reasoning_effort and verbosity, while new freeform tool-calling features expand tool compatibility beyond rigid schema constraints.
Orchestrate with the Model Router and Scale with Agents
The introduction of GPT-5 represents more than a model update; it's a platform evolution. The model router within Foundry Models intelligently selects the optimal GPT-5 variant or other suitable models based on prompt complexity, performance requirements, and cost efficiency, potentially saving up to 60% on inferencing costs without compromising fidelity. Anticipated soon, GPT-5 will be integrated with the Foundry Agent Service. This integration will combine frontier models with built-in tools like browser automation and Model Context Protocol (MCP) integrations. This will empower policy-governed, tool-using agents capable of web searches, interactions within web applications, and the completion of end-to-end tasks, all while ensuring telemetry, adherence to Microsoft Responsible AI principles, and robust security.Accelerating Business Impact with GPT-5
GPT-5's capabilities translate directly into tangible business value across various sectors:GPT-5 Customer Spotlight
SAP
"SAP is excited to be among the first to leverage the power of GPT-5 in Azure AI Foundry within our generative AI hub in AI Foundation. GPT-5 will enable our product team and developer community to deliver impactful business innovations to our customers."
—Dr. Walter Sun, SVP and Global Head of AI, SAP SE
Relativity
"GPT-5 in Azure AI Foundry raises the bar for putting legal data intelligence into action. This next-generation AI empowers legal teams to uncover deeper insights, accelerate decision-making, and drive stronger strategies across the legal process."
—Dr. Aron Ahmadia, Senior Director, Applied Science, Relativity
Hebbia
"The partnership between Hebbia and Azure AI Foundry gives financial professionals an unprecedented edge. GPT-5’s advanced reasoning helps pinpoint critical figures across thousands of documents and structure complex financial analysis with speed and accuracy."
—Danny Wheller, VP of Business and Strategy
Building with AI in GitHub Copilot and Visual Studio Code
Beginning today, GPT-5 is rolling out to millions of developers utilizing GitHub Copilot and Visual Studio Code. The model’s advanced reasoning capabilities support increasingly complex coding tasks, from sophisticated refactoring to navigating extensive codebases, thereby helping developers write, test, and deploy code with greater speed and quality. The latest VS Code release enhances the agentic coding experience with improved autonomous task handling, support for over 128 tools in chat requests, and chat checkpoints for restoring workspace changes. Furthermore, an updated Azure AI Foundry extension empowers developers to build agents directly within VS Code.Security, Safety, and Governance by Design
Security and safety are core tenets of GPT-5 and Azure AI Foundry:Start Building Today
GPT-5 is available through the Standard offering in Azure AI Foundry, with deployment options optimized for cost efficiency and governance, including Global and Data Zone (United States, European Union) choices for data residency and compliance. With Azure AI Foundry’s reliability, real-time evaluations, built-in observability, and secure deployment, organizations can confidently transition from pilot projects to full production. Unique tools like the Model Router further optimize quality, latency, and cost across various workloads.Learn more about Azure AI Foundry
Source: Originally published at Microsoft Azure Blog on August 7, 2025.