OpenAI launches GPT-Live-1 in the API for voice apps

OpenAI has introduced GPTLive1 in its API, giving developers a new voice model for building realtime voice apps and business workflows. The company says the model can listen and speak at the same time, with support for smoother interruptions, turn detection, and longer conversations. The release also adds more control over tone, pace, style, and tool delegation. OpenAI says developers can pair GPTLive1 with backend models for deeper reasoning and use it in settings such as customer support, reservations, and other phonebased tasks. OpenAI says GPTLive1 is available in the API now, with pricing set at $0.05 per minute for the frontend voice layer. The company also highlighted customer feedback and benchmark results showing improved turntaking and lower latency compared with earlier voice systems.