OpenAI Launches GPT-Live-1 Voice Models, Positioning Voice as Future Interface for AI Agentic Work

OpenAI unveiled new conversational AI models, GPT-Live-1 and GPT-Live-1 mini, designed to sound more natural and manage conversational turn-taking. The full-duplex models allow simultaneous speaking and listening, enabling users to interrupt naturally and supporting features such as live translation. GPT-Live-1 mini will replace the current Advanced Voice Mode in ChatGPT by default, with paid-tier users gaining access to the larger GPT-Live-1 model, which routes queries to OpenAI’s latest text models, including GPT-5.5, for search, reasoning and agentic tasks while maintaining conversation flow.

During a press briefing, ChatGPT Voice product lead Atty Eleti said the new mode is built for longer conversations, describing personal use during extended walks, and said OpenAI views voice as a potential primary interface for managing complex, long-running agentic work. The company did not provide details on rumored hardware products, including reported plans for AI-enabled earbuds. OpenAI said more than 150 million people currently use ChatGPT’s voice and dictation features, and noted that safeguards remain in place for age-appropriate responses and sensitive topics such as self-harm.

The announcement comes as competitors including Apple, Amazon and startups such as Sesame and Monogram also work to make AI assistants more conversational and visually interactive.