OpenAI's production-ready speech-to-speech model with WebRTC support, enabling natural voice conversations with low latency and barge-in capability.