OpenAI has unveiled GPT-Live, a cutting-edge voice AI system designed to revolutionize the way we interact with technology by making conversations more fluid. Unlike traditional voice assistants, GPT-Live employs a full-duplex architecture, enabling it to speak and listen simultaneously. This allows for more natural interactions, as the AI can engage with users by interjecting acknowledgments like “mhmm” or “yeah,” thereby reducing the awkward pauses commonly associated with digital communication.
GPT-Live is also equipped to handle more sophisticated requests that require web searches or intricate reasoning. In such cases, it can seamlessly delegate these tasks to a more robust AI model operating in the background, all while maintaining the flow of conversation with the user. Currently, these complex tasks are managed by GPT-5.5, and OpenAI plans to incorporate support for even more advanced models in future updates.
This announcement comes on the heels of OpenAI’s recent confirmation that its GPT-5.6 model series will soon be publicly available, pending the completion of a comprehensive cybersecurity review. The GPT-5.6 series, known for its flagship Sol model, also includes the Terra and Luna variants, each offering unique capabilities.
In an effort to make these innovations accessible, OpenAI has begun deploying two versions of the new voice models—GPT-Live-1 and GPT-Live-1 mini—to its ChatGPT users across the globe. Furthermore, the company intends to expand the reach of GPT-Live by offering it through its API, which will allow developers and businesses to integrate real-time voice AI capabilities into their applications.