Microsoft Foundry Accelerates Agentic Era with General Availability of Production-Ready AI Agent Platform

Microsoft is significantly advancing the adoption of artificial intelligence in enterprise environments with the general availability of Microsoft Foundry, a comprehensive platform designed to streamline the creation, deployment, governance, and distribution of AI agents. The platform’s latest updates, announced at Microsoft Build, aim to empower organizations to move beyond experimentation and into production with reliable, observable, and business-outcome-aligned AI systems. With over 100,000 organizations already leveraging Microsoft Foundry, and leading companies like Adobe, Telefónica, and Tata Consultancy Services running agents in production, this expansion signifies a pivotal moment in the operationalization of AI.

The core promise of Microsoft Foundry, as articulated during Microsoft Build, is to enable developers to build AI agents within their existing workflows, deploy them on trusted infrastructure, and seamlessly integrate them with end-users, all without the need to piece together disparate platforms. This vision has now transitioned from roadmap to reality with three key sets of updates that are now generally available. These advancements collectively integrate frontier AI models, robust production agent runtimes, enterprise-grade identity and security controls, and distribution capabilities across Microsoft 365 into a unified platform. This integrated approach is poised to dramatically reduce the complexity and time-to-value for organizations seeking to harness the power of AI agents.

Foundry: The Premier Platform for AI Agent Development and Deployment

Microsoft Foundry is positioned as an end-to-end solution for the entire AI agent lifecycle. It consolidates the critical capabilities organizations require to bring AI agents into production, structured across three fundamental pillars: Building, Running, and Governing. These pillars work in concert to provide a cohesive experience, enabling businesses to not only develop sophisticated agents but also to manage and scale them effectively.

The platform’s commitment to an integrated experience is evident in its developer-centric approach. Agent development commences within familiar environments such as GitHub Copilot and Microsoft Visual Studio (VS) Code. The Foundry Toolkit for VS Code, alongside the Foundry skill, facilitates the seamless deployment of these agents to the Foundry platform. Whether developers are utilizing the Microsoft Agent Framework, the GitHub Copilot SDK, or the Claude Agent SDK, Foundry serves as the definitive production destination. This emphasis on developer familiarity and existing toolchain integration is crucial for widespread adoption and efficiency.

Empowering Choice: Access to Frontier Models for Every Workload

A cornerstone of any capable AI agent is the underlying reasoning model. Microsoft Foundry addresses this by providing access to a diverse range of industry-leading frontier, open-source, and task-specific models through a single, unified platform. This allows teams to select the optimal model for each unique workload, ensuring both performance and cost-efficiency.

A significant development is the general availability of OpenAI’s GPT-5.6 series within Microsoft Foundry Models and Microsoft Foundry Agent Service. This series includes GPT-5.6 Sol, GPT-5.6 Terra, and GPT-5.6 Luna, each offering distinct levels of capability and cost. This tiered approach provides organizations with the flexibility to align model selection with specific business requirements, moving away from a one-size-fits-all model strategy.

The GPT-5.6 series is designed to offer a spectrum of performance and pricing, catering to a wide array of use cases. For instance, GPT-5.6 Sol is positioned as the most capable, likely offering advanced reasoning and complex task execution, while GPT-5.6 Luna represents a more cost-effective option for less demanding applications. GPT-5.6 Terra would bridge this gap, providing a balanced solution. This granular control over model choice is a critical factor for businesses managing AI budgets and optimizing operational expenses.

Customers have consistently emphasized the importance of timely access to the latest AI models, alongside model quality. Microsoft Foundry is addressing this by making GPT-5.6 available across 28 global regions from day one. This includes Global Standard and Global Priority Processing, Data Zones Standard, and Global Provisioned deployment options. This ensures that organizations can readily adopt cutting-edge AI innovations within their existing Azure infrastructure, facilitating seamless integration and scalability.

The pricing structure for the GPT-5.6 series within Microsoft Foundry further underscores this flexibility. For standard global deployments, GPT-5.6 Sol is priced at $5.00 per million input tokens and $30.00 per million output tokens. GPT-5.6 Terra is available at $2.50 per million input tokens and $15.00 per million output tokens, while the more economical GPT-5.6 Luna is priced at $1.00 per million input tokens and $6.00 per million output tokens. This tiered pricing allows businesses to strategically deploy different models for different tasks, optimizing cost while maintaining performance.

Global Reach and Data Sovereignty with Asia-Pacific Data Zone

Beyond model availability, the ability to run advanced AI compliantly is paramount. Microsoft Foundry is expanding its global footprint with the general availability of the Asia-Pacific (APAC) Data Zone. This new offering allows organizations in the APAC region to leverage frontier OpenAI models while ensuring that data processing remains within their geographical boundaries. This eliminates the need for separate, complex environments and accelerates the adoption of AI without compromising data sovereignty or compliance requirements.

The availability of Global, Data Zone, and Regional deployment options within Foundry provides organizations with unparalleled flexibility. They can align their AI strategies with critical considerations such as data sovereignty, regulatory compliance, performance demands, and scalability needs, all while maintaining a consistent development and operational experience across different deployment models.

Hongsoo Kim, Chief Data and AI Officer (CDAO) at Viva Republica (Toss), commented on the significance of the APAC Data Zone, stating, "As financial institutions adopt AI, responsible data handling becomes foundational to trust. Microsoft Foundry’s APAC Data Zone allows us to keep data processing regionally anchored while accessing advanced AI models at scale. This gives us the confidence to accelerate AI innovation responsibly and reinforces our ambition to be a leading AI-powered financial platform in Asia." This statement highlights the critical role of data governance in enterprise AI adoption and the value of localized data processing capabilities.

Building Impactful Agents: Action-Oriented and Context-Aware

A powerful model is only the initial component of a functional AI agent. To be effective in production, an agent requires a robust runtime environment, access to business-specific knowledge, governed access to tools, persistent memory across interactions, the ability to act on real-world events, and a clear pathway to end-users. Microsoft Foundry integrates these essential components as built-in capabilities designed to work seamlessly together.

The platform provides built-in capabilities for:

• Agent Runtime: A stable and scalable environment for agents to execute their functions.
• Tooling and Orchestration: Enables agents to interact with external systems and services through well-defined APIs.
• Knowledge Integration: Facilitates the incorporation of proprietary business data and external knowledge sources.
• Memory and State Management: Allows agents to retain context and learn from past interactions.
• Event-Driven Actions: Empowers agents to respond to real-time events and trigger automated workflows.
• Distribution Channels: Integrates agents with user-facing applications and workflows, including Microsoft 365.

This comprehensive set of features ensures that organizations can build agents that are not only intelligent but also actionable and deeply integrated into their business processes.

Governing and Optimizing the AI Lifecycle: Observability and Control

The ability to monitor, improve, and secure AI agents is critical for their successful deployment and long-term viability. Microsoft Foundry prioritizes trust as a core platform tenet, ensuring that these responsibilities do not fall solely on developers. The latest release introduces enhancements that address the post-build phase, providing visibility into agent performance, enabling continuous improvement, and demonstrating tangible value.

Key enhancements for governance and optimization include:

• Agent Observability: Comprehensive monitoring and logging capabilities provide insights into agent behavior, performance, and resource utilization. This allows teams to identify and troubleshoot issues proactively.
• Evaluation Frameworks: Tools and methodologies for assessing agent accuracy, reliability, and business impact, enabling data-driven improvements.
• Security and Compliance Controls: Enterprise-grade security features and compliance certifications ensure that agents operate within organizational security policies and regulatory requirements.
• Cost Management and Optimization: Granular controls and insights into AI spending, enabling teams to manage costs effectively without compromising performance.

As agents scale from pilot programs to handling thousands of daily requests, Foundry equips teams with the necessary tools to maintain predictable spending without leaving the platform. This cost management is facilitated through a multi-pronged approach:

• Model Routing and Optimization: The platform offers intelligent model routing, prompt caching, and efficient traffic management to minimize redundant computations and optimize resource allocation.
• Cost-Aware Agent Design: Tools like PTU spillover and quota optimization ensure service continuity during usage spikes, preventing unexpected cost overruns.
• Specialized Tooling: Toolboxes in Foundry ensure that agents only utilize the necessary tools for each request, reducing unnecessary processing.
• Agent Optimizer: This feature fine-tunes prompts, skills, tools, and model choices against custom evaluators, further enhancing efficiency and effectiveness.

Crucially, Foundry provides an integrated view of Return on Investment (ROI) for agents. This feature connects business value, usage metrics, and associated costs, allowing teams to clearly assess whether production agents are generating more value than they consume in resources. This transparency is essential for demonstrating the business impact of AI initiatives and for making informed decisions about future investments.

A Microsoft Mechanics episode on token economics for agents offers a detailed walkthrough of these cost management and ROI features.

Real-World Impact: Organizations Building on Foundry

The adoption of Microsoft Foundry extends beyond experimentation, with a growing number of organizations actively deploying AI agents in production. This includes a diverse range of entities, from digital-native startups to Fortune 500 enterprises. These companies are experiencing a significant acceleration in their AI initiatives, reducing the time required for integration, security, and deployment from weeks to days. By leveraging infrastructure that meets their stringent compliance standards and distributing agents through familiar user tools like Microsoft 365, these organizations are achieving tangible business outcomes more rapidly.

Getting Started with Microsoft Foundry

All the capabilities detailed in this announcement are currently live and accessible within Microsoft Foundry. Aspiring users can access comprehensive documentation and Microsoft Learn courses to guide their journey. Developers can begin their experience immediately by following the provided Quickstart guide, which offers an end-to-end walkthrough of setting up, testing, and deploying a production-ready hosted agent.

Further educational resources include:

• AI Agents for Beginners: A 12-lesson curriculum designed for foundational learning.
• Guided Labs: Including "Develop AI Agents in Azure," "Hosted Agents Workshop (.NET)," "Foundry Toolkit for VS Code and hosted agents workshop," and the "ZavaShop Supply Chain Workshop."
• Practical Guide: "Evaluating AI Agents: A Practical Guide with Microsoft Foundry" provides best practices for ensuring agent quality.

For a visual and in-depth understanding, the "Foundry Agent Service + Microsoft Agent Framework Explained" video by Jeff Hollan offers a detailed walkthrough of operationalizing AI agents from deployment to real-world impact. This array of resources underscores Microsoft’s commitment to empowering developers and organizations to successfully adopt and leverage AI agents.