Microsoft adds Anthropic’s AI to Copilot (techcrunch.com)

Microsoft is expanding Copilot’s backend beyond OpenAI by adding Anthropic’s models — Claude Opus 4.1 and Claude Sonnet 4 — to its enterprise Copilot lineup. The move follows a recent agreement to bring Anthropic into Office 365 apps and marks a broader shift away from a single-provider strategy. Business users will now be able to pick between OpenAI’s deep-reasoning models and Anthropic’s offerings for tasks like complex research, building custom AI tools, and assembling enterprise-grade agents. Technically, Opus 4.1 is positioned for high‑complexity workloads (deep reasoning, coding, and architecture planning), while Sonnet 4 is optimized for routine development, large-scale data processing, and content generation. That task-based model selection enables Microsoft to route requests to the model best suited for a job, potentially improving accuracy, cost-efficiency, and developer productivity. Strategically, the integration reduces vendor lock-in, heightens competition between major model providers, and gives enterprises more flexibility for compliance, safety, and performance trade-offs — though it also raises operational considerations around latency, monitoring, and model governance when orchestrating multiple LLMs.