Phi models you can run locally

Microsoft's Phi family focuses on capable small language models designed for efficient local and edge inference.