Microsoft has officially integrated Anthropic’s Claude Sonnet 4 and Opus 4 models into GitHub Copilot Chat, making the feature available to users of JetBrains IDEs and Eclipse.
On the GitHub platform, this AI coding agent has already taken on practical programming tasks, including bug fixes, feature additions, documentation enhancements, and test execution. GitHub CEO Thomas Dohmke noted that when developers assign tasks, the AI responds with emojis and starts working immediately—it creates new files within the same system environment to track outputs and provides change summaries, designed to act more like a “digital colleague” than an external tool.
This move marks a major shift in Microsoft’s AI strategy: moving from heavy reliance on its in-house models to adopting high-quality third-party AI solutions. The Claude 4 series has already gained support in Microsoft cloud services such as Azure Databricks, though due to high demand, some users need to submit a special request to activate the API endpoint for Claude 4 Sonnet.
AI Agent Strategy Becomes Core of Microsoft’s Productivity Vision
At the recently held Microsoft Build 2025 Developer Conference, Microsoft explicitly stated that AI agents will gradually become a core component of how people work, build software, and manage tasks. The company introduced the concept of the “open agentic web,” where AI agents can execute tasks and make decisions at both the individual and team levels.
Microsoft Chief Technology Officer Kevin Scott emphasized that these AI agents will not serve as background auxiliary tools, but rather exist as users’ “digital colleagues.”
Microsoft’s goal is to develop AI tools that operate within existing systems, rather than building independent AI application environments. GitHub Copilot—an early success powered by Claude Sonnet 3.7 —has proven the effectiveness of this integration model, paving the way for similar integrations with the Office suite in the future.
As AI agent capabilities advance, Microsoft has observed that users require greater depth and continuity of reasoning for complex tasks. The “deep and continuous reasoning” capability highlighted by the Claude 4 series is particularly well-suited for scenarios requiring analysis of dense documents or the generation of well-structured reasoning outputs—needs commonly seen in office environments.
Multi-Model Strategy Emerges as New Trend in Enterprise AI
Microsoft’s decision to adopt Claude 4 reflects an AI new trend in enterprise: moving away from reliance on a single vendor and instead selecting the most suitable model for specific tasks. Sources within Microsoft revealed that the Office 365 team is evaluating the possibility of integrating Claude Sonnet 4 into core applications such as Word, Excel, and Outlook—especially for scenarios requiring in-depth text analysis and complex workflow automation.
This strategic shift stems from a detailed assessment of the strengths of different models: while GPT-4 excels in response speed and versatility, Claude Sonnet 4 offers higher accuracy and reliability for tasks requiring long-context understanding and precise reasoning.
Notably, the Claude 4 series does not support image generation—a stark contrast to GPT-4. Microsoft may adopt a hybrid approach, continuing to use GPT models for image processing scenarios while integrating Claude 4 for text-intensive tasks, thereby building a more comprehensive AI-powered office experience.
