Microsoft 365 users can now choose between ChatGPT and Claude for their AI needs (www.techradar.com)

Microsoft has added Anthropic’s Claude Sonnet 4 and Claude Opus 4.1 as selectable models inside Copilot, introducing a “Try Claude” option in the researcher agent experience and making Claude available in Copilot Studio for building custom AI agents. The roll‑out is already appearing for Frontier program participants and opt‑in Microsoft 365 Copilot users; developers have been edging in this direction too (GitHub Copilot in VS Code recently started favoring Sonnet 4). This is a notable pivot away from Microsoft’s previous, near‑exclusive use of OpenAI models following OpenAI’s move to other cloud partners. Technically and strategically the shift matters: Anthropic’s models will be accessed via an API and hosted on AWS infrastructure rather than Azure, which raises new considerations for latency, data residency, compliance, and enterprise procurement. Early benchmarks suggest Claude performs strongly on Excel and PowerPoint tasks, so integrating it into Office workflows could change which backends enterprises prefer for productivity AI. For the AI/ML community the move accelerates multi‑model, multi‑cloud productization—encouraging competition, easier experimentation in Copilot Studio, and more choices for building specialized agents, while also forcing organizations to re‑evaluate architecture and governance around third‑party hosted models.