Microsoft is now offering OpenAI’s new free and open GPT model, GPT-OSS-20B, to Windows 11 users through Windows AI Foundry. This platform allows users to access AI features, APIs, and popular open-source models directly on their computers.
According to Microsoft, GPT-OSS-20B is designed to be lightweight and efficient, optimized for tasks like code execution and tool use. It runs smoothly on a variety of Windows hardware, with plans to expand support to more devices soon. The model is ideal for building autonomous assistants or integrating AI into real-world workflows, even in environments with limited bandwidth.
Launched recently, GPT-OSS-20B can operate on consumer PCs and laptops with at least 16GB of VRAM, which is supported by modern GPUs from Nvidia or Radeon. OpenAI trained the model using high-compute reinforcement learning, enhancing its ability to power AI agents and execute tasks like web searches or Python code as part of its reasoning process.
However, the model is text-only and cannot generate or process images and audio, unlike some of OpenAI’s other models. Another limitation is its tendency to hallucinate—OpenAI’s internal testing showed that GPT-OSS-20B produced incorrect responses to 53% of questions on its PersonQA benchmark, which measures factual accuracy about people.
Microsoft plans to bring GPT-OSS-20B to macOS and other devices in the future, though no specific details were provided. Additionally, both GPT-OSS-20B and the older GPT-OSS-120B are available through Microsoft’s Azure AI Foundry platform. These models are also accessible on Amazon’s AWS.
The rest of the content appears to be an unrelated promotional section about TechCrunch Disrupt 2025, featuring speakers from companies like Netflix, ElevenLabs, Wayve, and Sequoia Capital. The event will take place in San Francisco from October 27-29, 2025, with early ticket discounts available until August 7.