Microsoft has made a strategic decision to favor Anthropic's Claude Sonnet 4 over OpenAI's GPT-5 for its GitHub Copilot, citing superior performance in internal benchmarks. This shift extends to Microsoft 365 Copilot, where Anthropic's models will enhance various features. Julia Liuson, head of Microsoft's developer division, confirmed that Claude Sonnet 4 is the recommended model for GitHub Copilot users. Additionally, Microsoft is investing heavily in developing its own AI models, with plans for larger-scale training beyond its current capabilities. This decision comes despite Microsoft's significant financial commitment to OpenAI, which includes a $13 billion investment. The company is also preparing to update its Visual Studio Code editor to prioritize Anthropic's model, indicating a clear preference for Anthropic's technology in coding and development tasks.