ChatGPT Voice Mode Enhanced for Natural Conversation

OpenAI is overhauling ChatGPT's voice mode with a new model that aims to make interactions feel more like "talking to another person." The updated GPT-Live-1 is designed to interrupt users less and will wait for the user to continue speaking after a pause, enhancing conversational flow.

The new model is capable of recognizing when a user has finished speaking and can also interrupt its own response if necessary. Kundan Kumar, OpenAI's research lead, described GPT-Live-1 as the company's "smartest voice model yet."

GPT-Live-1 will automatically pass queries to its best text models, such as GPT-4o, for reasoning and web searches. This allows for a quicker transition from researching a topic to discussing its findings with the user. The model can also indicate when it is actively searching for information.

OpenAI has focused on improving the interactivity and naturalness of its AI, with advancements in voice mode being a key part of this effort. The enhanced voice capabilities are now available for users of the latest ChatGPT versions.