Phi-4 Reasoning

Big reasoning power, small models

Small open-weight models (3.8B/14B) delivering powerful reasoning for math/science/code, rivaling larger LLMs. Available on Azure AI Foundry & HF.