OpenAI Launches GPT-Live For Natural Real-Time Voice Conversations
OpenAI says its new voice system can listen, speak, and hand off harder work in the background without breaking the flow.
OpenAI says GPT-Live is a new generation of voice models now powering ChatGPT Voice, built around a full-duplex setup that can listen and speak at the same time. In a widely shared post on X, @kimmonismus described the pitch as more natural back-and-forth, quick verbal acknowledgments, and a background handoff to a heavier model for search or deeper reasoning while the conversation keeps moving. The same post says GPT-Live-1 and GPT-Live-1 mini are rolling out globally today, with API access coming soon.
OpenAI just launched GPT-Live, a new generation of voice models now powering ChatGPT Voice. The core shift is full-duplex.
Commentary on X
Highest ranked@athyuttamre It feels shaped by a shallow idea of what “natural” means: fake hesitations, exaggerated warmth, performed emotion. For poetry or philosophy, this breaks the experience. I don’t want a voice that tries to be cute. I want one that is sober, precise, and truly present:(