Phi-4 multimodal reasoning, scalable LLM systems, cinematic video generation, and AI for health

DRANK

In this issue: Phi-4-reasoning-vision-15B: A compact multimodal reasoning model for vision-language tasks, now available via Microsoft Foundry, HuggingFace, and GitHub Niyama (QoServe): QoS-driven scheduling system that boosts LLM serving capacity while reducing SLO violations NetArena: A dynamic be