Microsoft reportedly pushing Anthropic's Claude over OpenAI's GPT for coding assistance

In a significant and surprising shift in its AI strategy, Microsoft is reportedly prioritizing Anthropic's Claude over OpenAI's GPT for coding assistance within its popular Visual Studio Code editor. The move comes despite Microsoft's massive multi-billion-dollar investment in OpenAI, signaling a pragmatic focus on performance and a strategic effort to diversify its AI dependencies.

According to internal communications and industry reports, Microsoft has been quietly recommending Anthropic's models to its own developers for several months. In a recent internal email, Julia Liuson, the head of Microsoft's developer division, reportedly stated, “Based on internal benchmarks, Claude Sonnet 4 is our recommended model for GitHub Copilot.” This guidance, issued before the launch of OpenAI's GPT-5, has not changed.

Additionally, the new "auto model selection" feature for GitHub Copilot in Visual Studio Code, which dynamically chooses the best model for optimal performance, will primarily rely on Claude Sonnet 4 for paid users. This is a tacit admission that, for a number of coding tasks, Anthropic's models are currently outperforming OpenAI's.

Sources familiar with the matter indicate that the decision is driven by performance metrics. Anthropic's Claude models have reportedly demonstrated superior capabilities in tasks such as code generation, understanding complex context, and providing cleaner, more relevant suggestions.

This is not limited to coding, as reports suggest Anthropic's models have also outperformed OpenAI's in certain tasks within Microsoft 365 applications like Excel and PowerPoint.

This strategic pivot is a clear move by Microsoft to hedge its bets in the rapidly evolving AI landscape. While its partnership with OpenAI remains a cornerstone of its AI strategy, Microsoft is actively working to build a "multi-model moat" that reduces its reliance on a single provider. This includes not only partnerships with other AI leaders like Anthropic but also significant investments in developing its own proprietary AI models.

Broader implications

For the tech industry, the decision highlights that in the fast-paced world of AI, performance often trumps even the most deeply entrenched partnerships. For developers, this means better, more efficient tools as Microsoft orchestrates a competitive ecosystem of AI models to deliver the best possible user experience.