GPT-Live is OpenAI’s new generation of voice models powering ChatGPT Voice. Its full-duplex architecture continuously processes speech input while generating speech output, enabling more natural interruptions, pauses, listening cues, live translation, and delegation of harder work to stronger text models.
Features
- Full-duplex voice interaction
- Continuous listening and speaking
- Natural interruption and pause handling
- Live translation
- Background delegation for complex tasks
- ChatGPT Voice integration
Use Cases
- Real-time voice conversation
- Spoken translation
- Mobile voice assistant
- Voice collaboration across tasks
- Accessibility interaction
- Long-form voice experiences
FAQ
GPT-Live is OpenAI’s new generation of voice models powering ChatGPT Voice. Its full-duplex architecture continuously processes speech input while generating speech output, enabling more natural interruptions, pauses, listening cues, live translation, and delegation of harder work to stronger text models. Core capabilities include: Full-duplex voice interaction, Continuous listening and speaking, Natural interruption and pause handling.
Common scenarios include: Real-time voice conversation, Spoken translation, Mobile voice assistant.
