OpenAI launched GPT-Live-1, a full-duplex speech model in its API, designed to eliminate delays in AI voice interactions. Utilizing a single model for input and output processing, it enables human-like smooth conversations with immediate interruption handling. The model supports background reasoning, customizable tones, and robust performance in noisy environments. Priced at $0.05 per minute for the frontend speech layer, it has been adopted by partners like Yelp and Intercom to enhance natural, face-to-face-like conversational experiences in customer service and other applications.