OpenAI Introduces GPT‑Live Voice Upgrade to Make ChatGPT Speak Naturally

OpenAI launches GPT‑Live, a new voice interface that adds full‑duplex audio to ChatGPT, allowing the system to listen and speak simultaneously, creating smoother back‑and‑forth conversations without waiting for prompts across devices.

Full‑Duplex Voice means GPT‑Live can process user speech while generating its reply, cutting response latency by half compared with turn‑based models and enabling more natural dialogue flow for applications like virtual assistants and customer support.

Developers receive an API update that streams audio packets in real time, supporting multiple languages and voice tones, and OpenAI provides sample SDKs for integration into smartphones, smart speakers, and web platforms.

User Experience tests show conversational pauses drop from several seconds to under one second, and early adopters report higher satisfaction scores, suggesting GPT‑Live could set a new standard for interactive AI communication.