OpenAI released o1-mini, a distilled version of their GPT-4o model, aimed at cost-efficient reasoning. The o1-mini model is 10 times cheaper and 60% as capable as GPT-4o on a subset of tasks. Builders can use o1-mini for applications where full GPT-4o capabilities are not required, reducing costs. The release reflects OpenAI's focus on making AI more accessible and affordable.