OpenAI launches GPT-Live-1 voice models for natural conversations
Original: OpenAI releases new voice models for more natural live conversations
Why This Matters
Full-duplex voice AI signals a shift toward voice as a primary computing interface for complex, long-running tasks.
OpenAI released two new conversational AI models, GPT-Live-1 and GPT-Live-1 mini, on July 8, 2026. The full-duplex models support simultaneous speaking and listening, enabling natural interruptions, live translation, and longer conversations. ChatGPT's Advanced Voice Mode will default to GPT-Live-1 mini, with paid users accessing the larger model.
OpenAI announced GPT-Live-1 and GPT-Live-1 mini, full-duplex voice models that can speak and listen simultaneously — a departure from the previous pipeline that chained speech-to-text, a language model, and text-to-speech. The new architecture allows natural interruptions and live translation. ChatGPT's Advanced Voice Mode will be replaced by GPT-Live-1 mini for all users, while paid-tier subscribers gain access to GPT-Live-1. The models route queries to newer text models such as GPT-5.5 for search, reasoning, and agentic tasks without breaking the conversation flow. They can also remain silent, absorb conversational context, and present information in visual formats. ChatGPT Voice product lead Atty Eleti noted he has held 30- to 40-minute conversations during walks, and stated: "We think voice can be the future interface to all kinds of work." OpenAI reports over 150 million users engage with ChatGPT via Voice and Dictation features. Competitors including Apple, Amazon, and startups like Sesame are also advancing conversational AI assistants. No hardware product announcements were made.