OpenAI has launched a voice model that can talk and listen at the same time, enabling full-duplex conversations. The model, referred to as GPT-Live or GPT-Live-1 in various reports, is designed for real-time conversations, although some outlets suggest it may have additional capabilities such as thinking and reasoning like a human. The launch appears to be part of a broader release of voice models, with some reports indicating that it precedes a larger release of GPT-5.6.