OpenAI has introduced new models that can listen, translate, and act in real time. The company has unveiled multiple audio models for real-time voice tasks, with some reports specifying that there are three models. The models are part of an effort to enhance voice intelligence and may be integrated into the company's API for more natural conversations.