OpenAI rolls out new voice model for ChatGPT that listens and speaks simultaneously

OpenAI introduced a new voice model for ChatGPT that can speak and listen at the same time. The model, called GPT-Live-1, is designed to interrupt users less and to wait for them to continue speaking after a pause. The company said the model will automatically route complex queries to its strongest text models when reasoning or web searches are required.

It will also display AI-generated visuals such as sports scores or weather forecasts during relevant conversations.

Model capabilities OpenAI stated that GPT-Live-1 supports real-time translation while a user is still speaking. ” The model includes built-in safeguards that steer responses away from harmful content and end conversations in higher-risk situations. It is also trained to offer expert-vetted crisis helpline support for self-harm topics and age-appropriate responses for teens.

is rolling out on iOS, Android, and the web.

It will power voice features for Go, Plus, and Pro subscribers, while a smaller GPT-Live-1 mini version will serve free users by default.