OpenAI has unveiled GPT-Live, its latest advancement in voice AI technology, aimed at enhancing the naturalness of conversational interactions by enabling simultaneous speaking and listening capabilities. This innovation, based on a full-duplex architecture, allows the AI to interject with natural affirmations like “mhmm” or “yeah” as users speak, thus facilitating smoother and more immediate exchanges without the need for extended pauses.
When faced with more intricate queries that demand web searches or advanced reasoning, GPT-Live can effortlessly delegate these tasks to a more potent 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, with plans to incorporate even newer models in forthcoming updates.
This development comes on the heels of OpenAI’s announcement regarding its forthcoming GPT-5.6 model series, which is set to be publicly available following a rigorous cybersecurity review. The GPT-5.6 lineup includes the premier Sol model, alongside its Terra and Luna counterparts, each designed to offer unique functionalities and enhancements.
OpenAI is in the process of deploying two iterations of its new voice models—GPT-Live-1 and GPT-Live-1 mini—to ChatGPT users around the globe. In addition to this, the company intends to extend GPT-Live’s capabilities through its API, thereby empowering developers and businesses to integrate real-time voice AI into their applications.