What you'll learn

Hosted AI models are capable, but they come with some tradeoffs. You get charged per token, and you're sending your data to a third party. In this course, Introduction to OpenAI Open Weight Models, you'll learn to run and customize OpenAI's open-weight gpt-oss models on hardware you control. First, you'll explore how to choose between gpt-oss-20b and gpt-oss-120b and run them locally and in the cloud with tool use and structured output. Next, you'll discover how to tune reasoning effort and use the 128k context window for long-form and multi-document tasks. Finally, see an example of fine-tuning gpt-oss on a domain-specific dataset and how the result compares to frontier models. When you're finished with this course, you'll have the skills and knowledge of running and customizing gpt-oss needed to start building private, cost-effective AI applications.