OpenAI releases new voice models for more natural live conversations

OpenAI on Wednesday released new conversational voice models called GPT-Live-1 and GPT-Live-1 mini that the company says sound more natural and handle turn-taking better than previous versions. The models are full-duplex, allowing them to speak and listen at the same time so users can interrupt naturally and enabling live translation, according to TechCrunch. OpenAI is replacing its current Advanced Voice Mode in ChatGPT with GPT-Live-1 mini by default for all users, while paid tier users can access the larger GPT-Live-1 model. The previous system used separate speech-to-text, language model and text-to-speech components. The new models route queries to OpenAI's latest text models such as GPT-5.5 for reasoning and other capabilities while maintaining the conversation. In a press briefing, ChatGPT Voice product lead Atty Eleti said he has had conversations lasting 30 to 40 minutes during walks. Eleti said the models allow the assistant to stay silent for extended periods while absorbing context. More than 150 million people use ChatGPT voice and dictation features, the company said. The announcement was reported by Ivan Mehta at TechCrunch.