Microsoft is enhancing on-device AI by introducing OpenAI’s gpt-oss models to local PCs through its new Azure AI Foundry Local platform. This major initiative enables developers to run sophisticated AI applications on Windows and macOS without an Azure subscription, ensuring privacy, security, and low-latency interactions. Foundry Local allows offline processing, utilizing the ONNX Runtime for optimized performance across various hardware. Supporting both gpt-oss-20b and gpt-oss-120b models, Foundry Local aims to meet the increasing demand for locally-run AI solutions. These models, built on a Mixture-of-Experts architecture, promote efficiency while minimizing computational costs. The launch aligns with strategic partnerships, particularly with Qualcomm and AWS, expanding OpenAI’s reach. This collaborative effort signifies a shift towards democratizing AI workloads and setting new standards in the open-source community, positioning Microsoft and OpenAI as leaders in the evolving AI landscape.
