OpenAI has unveiled GPT-Live, a cutting-edge voice AI technology aimed at enhancing the fluidity of human-machine interactions by enabling simultaneous speaking and listening. Leveraging a full-duplex architecture, GPT-Live can engage in conversations without delay, offering natural acknowledgments like “mhmm” or “yeah,” which help maintain the flow of dialogue without awkward pauses.
For more intricate tasks requiring web searches or sophisticated reasoning, GPT-Live seamlessly delegates these to a more advanced AI model, ensuring the conversation continues uninterrupted. At its launch, the model handling these tasks is GPT-5.5, with plans to incorporate newer models in upcoming updates.
This development follows OpenAI’s recent confirmation that its forthcoming GPT-5.6 models, including the Sol, Terra, and Luna variants, will be publicly released after a thorough cybersecurity review. The GPT-5.6 series represents the next stage in AI evolution, promising enhanced capabilities upon its approval.
OpenAI has already begun the global rollout of two iterations of the new voice models, known as GPT-Live-1 and GPT-Live-1 mini, to ChatGPT users. Additionally, the company intends to offer GPT-Live through its API, allowing developers and businesses to integrate real-time voice AI features into their applications.