OpenAI Introduced New Voice Model GPT-Live for ChatGPT

OpenAI has introduced GPT-Live, a new voice model for ChatGPT that makes conversations with users more natural and engaging.

The new system uses full-duplex technology, which means it can listen to you and respond simultaneously. During a conversation, GPT-Live reacts naturally, making brief comments when necessary, responding quickly, or simply waiting while you continue speaking.

Unlike previous versions, GPT-Live processes information received during the conversation instantly and decides multiple times per second how to respond to the user. This makes the dialogue much smoother and more natural. If a user requests a more complex task, such as searching for information on the internet or conducting in-depth analysis, the system uses the GPT-5.5 model.

Users can also determine the speed at which the model operates. Instant is designed for quick responses, while Medium and High are intended for more complex questions and tasks.

GPT-Live has been released in two versions. GPT-Live-1 is available to users of the Go, Plus, and Pro plans, while GPT-Live-1 mini can be used by free-tier users. The feature is already working in iOS and Android applications, as well as on the web version of ChatGPT.

During conversations, GPT-Live can also display additional information on the screen. For example, weather forecasts, stock prices, or sports news. At this stage, the model does not yet support video calls and screen sharing, though OpenAI states it is already working on adding these capabilities.

Source: Investing.com