On-Device AI: Microsoft Launches OpenAI’s gpt-oss Models for Windows

Microsoft has announced a significant advancement in on-device artificial intelligence by bringing OpenAI’s new gpt-oss models directly to Windows users. The initiative, leveraging a new platform called Azure AI Foundry Local, allows developers to run powerful AI models locally on their PCs, enabling secure, private, and offline AI development. This marks OpenAI’s first open-weight model release since GPT-2, making advanced AI capabilities more accessible. The models, specifically gpt-oss-20b and gpt-oss-120b, are optimized to run on various hardware, including CPUs, GPUs, and NPUs, without requiring an internet connection or a connection to Microsoft’s cloud services. This move is designed to empower developers and businesses to create new applications with low latency and enhanced data security, as all processing is handled on the local machine. The integration of these models into the Windows ecosystem is a strategic step for Microsoft, further solidifying its position in the competitive and rapidly evolving field of decentralized AI.