Microsoft has taken a major leap in cloud AI by integrating Anthropic’s Claude models into Microsoft Foundry, creating a unique space where Azure users have access to both Claude and GPT models. This powerful combination gives businesses more flexibility and choice, driving faster and smarter innovation across industries.
What Sets Claude Models Apart?
The Claude family (Haiku 4.5, Sonnet 4.5, and Opus 4.1) caters to a variety of enterprise needs. Each version brings distinct strengths:
- Claude Haiku 4.5: Optimized for speed and cost, ideal for real-time and basic automation tasks.
- Claude Sonnet 4.5: Specializes in complex agent logic, advanced coding, and cybersecurity scenarios.
- Claude Opus 4.1: Excels at deep reasoning, long-form tasks, and sophisticated content generation.
All models are built with Constitutional AI for robust safety and benefit from Foundry’s enterprise-level governance and seamless integration—key for secure, scalable deployment.
Accelerating the Shift to Intelligent Agents
Organizations are moving from static applications toward dynamic, agentic AI systems. Microsoft Foundry simplifies this transition with a unified, secure environment for model testing, deployment, and scaling. Its robust governance structure helps companies safely operationalize AI agents, with the potential to improve efficiency by up to 30%.
Claude in Foundry Agent Service: Smarter Automation
Within Foundry Agent Service, Claude models serve as the reasoning engines behind dynamic, goal-driven agents. Their main capabilities include:
- Multi-step workflow orchestration for tackling complex, long-context operations
- Direct integration with productivity tools through the Model Context Protocol, streamlining data access and automation
- Automated data operations such as summarization, classification, and extracting insights
- Real-time model selection to balance latency and costs with Foundry’s model router
- Centralized controls and observability for managing agent operations across departments
These features empower developers to build agents that autonomously resolve issues, execute code, analyze data, and interact with enterprise systems—all within a secure framework.
Modular Skills: Reusable Building Blocks for Automation
Claude’s Skills system allows developers to define and reuse modular building blocks that combine:
- Natural-language instructions
- Optional Python or Bash code
- Connected data files or APIs
With Skills, teams can automate reporting, clean data, or transform content, and every Skill is governed, version-controlled, and traceable in Foundry, making best practices easy to share and scale.
Deep Research: Beyond Simple Responses
Claude’s Deep Research capability pushes beyond static answers. Agents can pull live data, cross-reference internal sources, and deliver grounded, analytical insights. This turns agents into strategic partners, synthesizing trends and evidence from wide-ranging datasets for real business impact.
Transparent, Usage-Based Pricing
Claude models are globally available in East US 2 and West US Azure regions, with clear, usage-based pricing:
- Haiku 4.5: $1.00/input million tokens, $5.00/output million tokens
- Sonnet 4.5: $3.00/input million tokens, $15.00/output million tokens
- Opus 4.1: $15.00/input million tokens, $75.00/output million tokens
The Future: Responsible AI at Scale
This partnership between Microsoft and Anthropic is about more than just new technology—it’s about empowering organizations to build with responsible, scalable AI. With Claude models in Foundry, businesses can unlock advanced agents, streamline operations, and drive new levels of productivity. Now is the time for developers and enterprises to leverage Claude’s intelligence and help shape the next generation of AI solutions on Azure.
Anthropic’s Claude Models Arrive in Microsoft Foundry: Azure’s New Frontier for Intelligent AI Agents