OpenAI Releases Voice Model That Can Think, Reason, Respond Like A Human

OpenAI has released GPT-Live, a new voice model for its ChatGPT platform. The model uses a full-duplex architecture, enabling it to listen and speak simultaneously, according to the company. OpenAI claims that GPT-Live can think and reason in real time while generating responses, marking a step toward more natural conversational AI.

MediaPost reported the launch, highlighting that the model represents a new generation of voice capabilities. Full-duplex communication allows the AI to process incoming audio while producing output, reducing pauses in conversation. The model is designed to be more responsive and human-like in its interactions.

The company continues to expand its voice features, with GPT-Live representing a significant upgrade in real-time interaction. No specific release date or pricing was mentioned in the initial announcement.