OpenAI has introduced GPT-Live, a new voice AI model designed for more natural, real-time conversations. Built on a full-duplex architecture, it can listen and speak simultaneously, respond with natural acknowledgments and smoothly interject during discussions.
GPT-Live can hand off complex tasks like web searches or reasoning to models such as GPT-5.5 while continuing the conversation, preventing interruptions. It also better understands pauses, filters out background noise and supports rich visual responses for topics like weather, sports and stocks.
OpenAI is launching GPT-Live-1 for ChatGPT Pro, Plus and Go users, and GPT-Live-mini for free users. The new voice experience is available starting today on ChatGPT’s mobile apps and web version.
Source: Android Authority
