Microsoft Integrates OpenAI’s GPT Model into Windows AI Foundry
OpenAI unveiled a new lightweight and open GPT model on August 6, 2025, which can run on personal computers, marking a significant step in accessible AI technology. Microsoft has swiftly facilitated this by making the gpt-oss-20b model available through its Windows AI Foundry, with a macOS version set to follow shortly, reports 24brussels.
The gpt-oss-20b model requires a PC or laptop equipped with at least 16GB of VRAM, necessitating high-performance GPUs from Nvidia or select Radeon models. This model is optimized for executing code and utilizing tools, positioning it as an ideal solution for developing autonomous assistants or integrating AI into practical workflows, even in bandwidth-restricted settings.
Microsoft has tailored the gpt-oss-20b model for local inference and has indicated that support for additional devices will be introduced soon. This could lead to an upgraded version suitable for Copilot Plus PCs, similar to Microsoft’s recent enhancements to local AI models within Windows.
The rapid integration of OpenAI’s latest model into the Windows platform coincides with Amazon’s proactive adoption of the new open-weight GPT-OSS models for its cloud services. This marks a unique occurrence where OpenAI’s model can be operated locally on Windows and simultaneously provides Microsoft’s main cloud rival access to cutting-edge OpenAI technologies, further complicating the existing dynamics of the OpenAI and Microsoft partnership.
