ChatGPT’s enhanced voice feature is more effective at muting interruptions.

OpenAI is revamping ChatGPT’s voice feature with a new model that aims to create a more natural conversational experience, akin to chatting with a friend. The latest iteration, GPT-Live-1, is designed to minimize interruptions and can wait for you to resume speaking if you pause during a discussion.

During a recent briefing, OpenAI’s research lead, Kundan Kumar, referred to GPT-Live-1 as the company’s most advanced voice model to date. This model intelligently routes your questions to its top text models, like GPT-5.5, whenever it needs to reason or look up information online. This allows for a smoother transition from researching your query to discussing the results. Additionally, the upgraded model enhances conversations about topics like weather, stocks, and sports by incorporating AI-generated visuals that display relevant data, such as sports scores or the week’s weather forecast.

Previously, ChatGPT utilized an older, turn-based voice model that sometimes struggled with accuracy and maintaining a fluid conversational style. “This is a full duplex model,” explained OpenAI product lead Atty Eleti during the briefing. “It can speak and listen simultaneously, processing input and generating output continuously.”

This new setup introduces exciting features for ChatGPT’s voice mode, including real-time translation. Now, instead of waiting for you to finish speaking, ChatGPT can translate your words as you talk. You can also instruct ChatGPT Voice to pause until prompted, a feature that wasn’t available before. It will acknowledge that it’s listening with responses like “mhmm,” “yeah,” and “got it.”

OpenAI has implemented built-in safeguards to prevent harmful responses and to end conversations in “higher-risk” scenarios. The model is also trained to provide “expert-vetted crisis helpline support” for discussions around self-harm and is designed to deliver “age-appropriate” responses for younger users.

GPT-Live-1 is being rolled out across iOS, Android, and the web. It will enhance ChatGPT Voice for Go, Plus, and Pro subscribers, while a more compact version, GPT-Live-1 mini, will serve as the default for free users.