According to Satya Nadella, Microsoft has made a new Multi-Model Intelligence capability available in Frontier, linking to Microsoft Tech Community’s Microsoft 365 Copilot blog. According to Microsoft Tech Community, the Researcher experience now orchestrates multiple foundation models—such as Microsoft’s in-house Phi family alongside third‑party large language models like GPT—to improve retrieval, synthesis, and citation for enterprise research workflows. As reported by Microsoft Tech Community, the system routes tasks to the best model for summarization, grounded search with Microsoft Graph, and source attribution, targeting lower latency and cost for routine queries via smaller models while escalating complex tasks to larger models. According to Microsoft Tech Community, business users can leverage this multi-model pipeline inside Microsoft 365 environments, enabling secure data grounding, traceable citations, and policy compliance, which creates opportunities to reduce research time, improve content quality, and optimize compute spend across departments.

According to God of Prompt on Twitter, recent testing shows that the Phi and Mistral models experienced a significant 13% accuracy drop when evaluated on the GSM1k benchmark compared to GSM8k. Some model variants saw drops as high as 13.4 percentage points. The analysis suggests these models are not demonstrating true reasoning abilities but rather memorization, as they were exposed to the correct answers during training. This finding highlights critical concerns about the generalization and reliability of these AI models for business and research applications.