At Microsoft Build 2024, new additions to the Phi-3 model family were unveiled, emphasizing advancements in multimodal capabilities and accessibility on Microsoft Azure. Phi-3-vision, a 4.2B parameter model combining language and vision capabilities, was introduced alongside the availability of Phi-3-small and Phi-3-medium on Azure, offering robust solutions for generative AI applications. These models are optimized for diverse hardware and can be deployed efficiently across various platforms. Developed with Microsoft’s responsible AI standards, Phi-3 models prioritize safety and performance, catering to a wide range of use cases from healthcare to educational tools.
