Microsoft's Visual Studio Code Prioritizes Anthropic's Claude 4 Over OpenAI's GPT-5
Microsoft's development environment, Visual Studio Code, has implemented a new AI model selection feature that leans towards Anthropic's Claude Sonnet 4 as the preferred option for enhanced performance in coding tasks. This decision highlights Microsoft's evolving AI strategy, which includes significant internal investments in AI model development and expanded collaborations beyond its long-standing partnership with OpenAI.
Microsoft Shifts AI Model Preference in Visual Studio Code, Embracing Anthropic's Claude 4
Redmond, Washington – On September 16, 2025, Microsoft unveiled a significant update to its popular code editor, Visual Studio Code. The new feature introduces an automated AI model selector designed to optimize performance for developers. Notably, this system prioritizes Anthropic's Claude Sonnet 4 over OpenAI's GPT-5 and GPT-5 mini for GitHub Copilot users. While free users will experience a dynamic selection across various models, paid subscribers will predominantly utilize Claude Sonnet 4.
This strategic pivot reflects a growing internal consensus within Microsoft. According to sources familiar with the company's developer initiatives, Microsoft has been directing its own developers to favor Claude Sonnet 4 in recent months. Julia Liuson, who leads Microsoft's developer division, articulated this preference in a June internal communication, stating that Claude Sonnet 4 is the recommended model for GitHub Copilot based on internal performance benchmarks. This guidance remains unchanged even after the release of GPT-5, indicating a clear preference for Anthropic's offerings in specific development contexts.
Beyond this integration, Microsoft is also making substantial investments in developing its proprietary AI models. Mustafa Suleyman, Microsoft's AI chief, revealed in a recent employee town hall that the company's MAI-1-preview model, despite being trained on a relatively modest cluster of 15,000 H100 GPUs, represents a fraction of Microsoft's broader ambitions in AI training. This indicates a dual-pronged approach: leveraging external partnerships while simultaneously strengthening in-house AI capabilities.
The collaboration with Anthropic extends beyond Visual Studio Code. Reports from The Information suggest that Microsoft is poised to integrate Anthropic's AI models into select features of its Microsoft 365 applications. This move comes after internal evaluations revealed that some Anthropic models delivered superior performance compared to OpenAI's counterparts in applications like Excel and PowerPoint.
These developments occur in the wake of a renewed agreement between OpenAI and Microsoft, announced just last week, which may facilitate OpenAI's initial public offering. Despite having invested over $13 billion in OpenAI since 2019 and maintaining a complex revenue-sharing arrangement, Microsoft's willingness to diversify its AI model portfolio and allow OpenAI to utilize rival cloud providers signals a strategic realignment. Further details regarding the next phase of the OpenAI partnership are anticipated soon, suggesting a potentially redefined relationship in the evolving AI landscape.
This shift in Microsoft's strategy highlights the intensely competitive and rapidly evolving nature of the AI industry. It underscores the importance for major tech companies to maintain flexibility and explore multiple AI partnerships and internal development initiatives to ensure optimal performance and innovation across their product ecosystems. For developers and users of Microsoft products, this could translate into access to a broader range of high-performing AI tools tailored for specific tasks, potentially fostering greater efficiency and new capabilities.