Phi-3: Microsoft’s small language model will make AI tools accessible for developers

Microsoft Phi-3 model

At the Build 2024 conference, Microsoft announced a host of new tools for developers. Among those is a new multimodal model in the the Phi-3 family of AI small language models (SLMs), developed by Microsoft. According to Jessica Hawk, corporate vice President, Data, AI, and Digital Applications, Product Marketing, Microsoft, “Phi-3 models are powerful, cost-effective and optimised for resource constrained environments including on-device, edge, offline inference, and latency bound scenarios where fast response times are critical.”

According to Hawk, the model offers the ability to input images and text, and to output text responses. Tiny but mighty is Microsoft’s idea for these models as they are sized at 4.2 billion parameters and support general visual reasoning tasks and chart/graph/table reasoning. “For example, users can ask questions about a chart or ask an open-ended question about specific images,” Hawk said. Phi-3-mini and Phi-3-medium are now generally available as part of Azure AI’s MaaS offering.

Microsoft claims that Phi-3-vision is the first multimodal model in the Phi-3 family, bringing together text and images, and the ability to reason over real-world images and extract and reason over text from images. It has also been optimised for chart and diagram understanding and can be used to generate insights and answer questions.

Microsoft wants to expand and simplify access to the AI, data—application platform services to help developers build new AI experiences. Hawk cited an example of how developers are using Phi-3. She pointed out ITC in India which has built a copilot for Indian farmers to ask questions about their crops in their own vernacular