OpenAI brings GPT-Live-1 full-duplex voice model to the API

OpenAI

Audio official + media 2 src. ~1 min

OpenAI released GPT-Live-1 in the API on September 10, bringing the full-duplex voice system behind ChatGPT Voice to developers. The model listens and speaks simultaneously, handles interruptions, and can delegate reasoning or actions to a backend agent, with custom voices and telephony support at $0.05 per minute.

Why it matters

First time OpenAI's realtime voice stack is exposed to third-party developers at commodity per-minute pricing.

Importance: 3/5

Notable API release exposing OpenAI's realtime voice stack; official + media

Sources