Microsoft’s Phi-4-multimodal AI model handles speech, text, and video

Microsoft has introduced two new small language models, Phi-4-multimodal and Phi-4-mini, designed for on-device AI applications. Phi-4-multimodal, a 5.6 billion parameter model, can process speech, vision, and text simultaneously using less computing power. It's suitable for use in smartphones, cars, and lightweight enterprise applications. The models aim to expand possibilities for developers creating AI apps for resource-constrained devices, though some experts note limitations for certain generative AI use cases on mobile devices.