Right this moment, we’re asserting common availability of OpenAI’s new flagship, GPT-5, in Azure AI Foundry. That is greater than a brand new mannequin launch; it is probably the most highly effective giant language mannequin (LLM) ever launched throughout key benchmarks.
For enterprise leaders constructing with AI, the dialog has moved past chat. The bar is greater: can your AI generate, motive, and ship measurable outcomes—safely and at scale?
Right this moment, we’re asserting common availability of OpenAI’s new flagship mannequin, GPT-5, in Azure AI Foundry. That is greater than a brand new mannequin launch; it is probably the most highly effective LLM ever launched throughout key benchmarks. GPT-5 in Azure AI Foundry pairs frontier reasoning with high-performance technology and price effectivity, delivered on Microsoft Azure’s enterprise-grade platform so organizations can transfer from pilots to manufacturing with confidence.
GPT-5 in Azure AI Foundry: Constructed for real-world workloads
In Azure AI Foundry, the GPT-5 fashions can be found through API and orchestrated by themodel router. The GPT-5 sequence spans complementary strengths:
- GPT-5, a full reasoning mannequin supplies deep, richer reasoning for analytics and sophisticated duties, like code technology, with a 272k token context.
- GPT-5 mini powers real-time experiences for apps and brokers that require reasoning, device calling to unravel buyer issues.
- GPT-5 nano is a brand new class of reasoning mannequin which focuses on ultra-low-latency and velocity with wealthy Q&A capabilities.
- GPT-5 chat allows pure, multimodal, multi-turn conversations that stay context-aware all through agentic workflows, with 128k token context.
Collectively, the suite delivers a seamless continuum from rigorous agentic coding duties, to comparatively easy Q&A—all delivered with the identical Azure AI Foundry endpoint utilizing mannequin router in Foundry Fashions.
Underneath the hood, GPT-5 unifies superior reasoning, code technology, and pure language interplay. It combines analytical depth with intuitive dialogue to unravel end-to-end issues and clarify its method. Agentic capabilities enable multi-step device use and lengthy motion chains with clear, auditable selections. As a frontier-level coding mannequin, GPT-5 can plan complicated agentic workflow, construct migrations, and refactor code, in addition to produce exams and documentation with clear rationale. Developer controls—together with parameters like reasoning_effort and verbosity—let groups tune depth, velocity, and element, whereas new freeform tool-calling options allow broadens device compatibility with out inflexible schemas.
Orchestrate with the mannequin router—then scale with brokers
Introducing GPT-5 to Azure AI Foundry is greater than a mannequin drop: it’s a leap ahead for the platform. Beginning right this moment, builders can use the mannequin router in Foundry Fashions to maximise the capabilities of the GPT-5 household fashions (and different fashions in Foundry Fashions) whereas saving as much as 60% on inferencing price with no loss in constancy. Powered by a fine-tuned SLM underneath the hood, the mannequin router evaluates every immediate and decides the optimum mannequin primarily based on the complexity, efficiency wants, and price effectivity of every activity. Let the mannequin router choose the correct mannequin so as to construct your AI-powered functions with ease.
And orchestration doesn’t cease at routing—Foundry carries the identical intelligence into brokers. Coming quickly, GPT-5 might be obtainable within the Foundry Agent Service, pairing frontier fashions with built-in instruments together with new browser automation and Mannequin Context Protocol (MCP) integrations. The outcome: policy-governed, tool-using brokers that may search, act in net apps, and full end-to-end duties—instrumented with Foundry telemetry and aligned to Microsoft Accountable AI.
Accelerating enterprise impression with GPT-5
These capabilities map on to enterprise impression.
In analysis and information work, GPT-5 accelerates monetary and authorized evaluation, market intelligence, and due diligence—studying at scale and producing decision-ready output with traceability. In operations and decisioning, it strengthens logistics help, threat evaluation, and claims processing by pairing sturdy reasoning with coverage adherence. Copilots and buyer expertise groups profit from multi-turn, multimodal brokers that motive in actual time, name instruments, resolve duties, and revert to people with extra useful context.
In software program engineering, GPT-5 excels at code technology, software modernization, and high quality engineering—enhancing code fashion and explanations to compress assessment cycles.
And to be used circumstances that are price or latency delicate, GPT-5-nano’s extremely‑low‑latency structure delivers speedy, excessive‑accuracy responses, making it the perfect goal for effective‑tuning and the go‑to mannequin for prime‑quantity, simple requests.
GPT-5 buyer highlight
Prospects are unleashing GPT-5 throughout complicated, mission-critical workloads—accelerating decision-making, supercharging coding, and catalyzing product innovation.
SAP
SAP is worked up to be among the many first to leverage the ability of GPT-5 in Azure AI Foundry inside our generative AI hub in AI Basis. GPT-5 in Azure AI Foundry will allow our product crew and our developer group to ship impactful enterprise improvements to our prospects.
—Dr. Walter Solar, SVP and World Head of AI, SAP SE
Relativity
The GPT-5 in Azure AI Foundry raises the bar for placing authorized knowledge intelligence into motion… This next-generation AI will empower authorized groups to uncover deeper insights, speed up decision-making, and drive stronger methods throughout the complete authorized course of.
—Dr. Aron Ahmadia, Senior Director, Utilized Science, Relativity
Hebbia
The partnership between Hebbia and Azure AI Foundry provides monetary professionals an unprecedented edge. With GPT-5’s superior reasoning in Hebbia, they’ll pinpoint vital figures throughout hundreds of paperwork and structuring complicated monetary evaluation with velocity and accuracy.
—Danny Wheller, VP of Enterprise and Technique
Constructing with AI in GitHub Copilot and Visible Studio Code
GPT-5 begins rolling out right this moment to tens of millions of builders utilizing GitHub Copilot and Visible Studio Code, making use of the flagship mannequin’s superior reasoning capabilities to more and more complicated issues—from refined refactoring to navigating giant codebases extra successfully. GPT-5 helps builders write, take a look at, and deploy code quicker, whereas supporting agentic coding duties with vital enhancements to coding fashion and general code high quality. With GPT-5, builders not solely code quicker, however code higher.
With right this moment’s VS Code launch, builders additionally acquire a extra highly effective agentic coding expertise immediately throughout the editor: GitHub Copilot’s coding agent has an improved expertise for autonomously tackling duties within the background. Moreover, the GitHub Copilot chat expertise brings elevated productiveness, together with help past 128 instruments for a single chat request and chat checkpoints permitting customers to revive workspace modifications to a previous level. Right this moment, we’re additionally asserting an up to date extension to develop brokers utilizing the Azure AI Foundry extension all inside VS Code setting.
These bulletins lengthen Microsoft’s technique to rework software program improvement with AI, bringing superior AI capabilities to the complete software program lifecycle.
Safety, security, and governance by design
In all domains, safety and security is a layer cake of protections, which collectively present safety for threat eventualities—and AI is not any totally different. For AI, we take into consideration layers with the mannequin because the core. With GPT-5, the core is safer than earlier than:
The Microsoft AI Purple Group discovered GPT-5 to have one of many strongest security profiles of any OpenAI mannequin, acting on par with—or higher than—o3.
—Dr. Sarah Hen, Chief Product Officer of Accountable AI, Microsoft
As we take into consideration the security, safety, and governance layers round this core—Azure AI Foundry supplies quite a lot of extra controls:
- Azure AI Content material Security protections are utilized to each immediate and completion, corresponding to immediate shields, which assist to detect and mitigate prompt-injection makes an attempt earlier than they attain the mannequin.
- Constructed-in agent evaluators work with the AI Purple Teaming Agent to run alignment, bias, and safety exams all through improvement and manufacturing, whereas steady analysis streams real-time metrics—latency, high quality, security, and equity—stream into Azure Monitor and Software Insights for single-pane visibility.
- Lastly, safety alerts combine immediately with Microsoft Defender for Cloud, and runtime metadata and analysis outcomes are built-in to Microsoft Purview for audit, data-loss prevention, and regulatory reporting, extending safety and governance throughout the complete GPT-5 lifecycle.
Bringing AI into each workflow with GitHub Copilot and Visible Studio Code
Beginning right this moment, GPT-5 begins rolling out to tens of millions of builders who use GitHub Copilot and Visible Studio Code who will be capable to choose GPT-5 to put in writing, take a look at, and deploy code—and develop brokers utilizing the Azure AI Foundry extension all inside VS Code setting. GPT-5 helps complicated agentic coding duties with vital enhancements to coding character, front-end aesthetics, and code high quality, extremely desired enhancements for the developer group.
Our evaluations present OpenAI GPT-5’s reasoning capabilities and contextual consciousness exceed o3, enabling builders to sort out extra complicated issues—from refactoring to navigating giant codebases. With GPT-5, customers within the Visible Studio household can’t solely code quicker, however code higher.
VS Code and our latest resolution to open-source GitHub Copilot, represents our dedication to open instruments and requirements and demonstrates our means to fulfill the speedy tempo of mannequin improvements whereas protecting the developer expertise on the forefront. In right this moment’s launch of VS Code, builders can. In right this moment’s VS Code launch, builders have much more management over their expertise in chat—with enhancements to the reliability of terminal instruments, updates to the device picker and limits, new checkpoints, and extra.
Right this moment’s announcement extends Microsoft’s technique to rework software program improvement with AI, bringing superior AI capabilities to the complete software program lifecycle.
Begin constructing right this moment
GPT-5 is offered through our Normal providing in Azure AI Foundry, with deployment decisions optimized for cost-efficiency and governance wants, together with World and Information Zone (United States, European Union) deployment choices for knowledge residency and compliance.1
With Azure AI Foundry’s first-class reliability, realtime evaluations, built-in observability, and safe deployment choices, you may confidently transfer from pilot to manufacturing—all aided whereas distinctive instruments like Mannequin Router optimizes high quality, latency, and price throughout workloads.
1Pricing is correct as of August 2025