OpenAI on Sept. 10, 2026, launched GPT-Live-1 in the API, bringing ChatGPT's full-duplex voice, which listens and speaks at the same time, to developer apps and business workflows.
Company claim: a single model handles interruption-aware conversation and avoids the latency of chained speech-to-text, text-model and text-to-speech stacks, while developers can still hand hard reasoning and tools to a backend such as GPT-6 Astra.
OpenAI says developers can steer tone, pace and style; the model handles background noise and silence more quietly; long sessions keep context better; and telephony support enables phone agents.
Partner quotes in the launch post claim large cuts in unwanted interruptions or voice-stack code. Those lines are partner testimonials, not FrameSignal measurements.
OpenAI says GPT-Live-1 is available now at $0.05 per minute for the front-end voice layer. Vendor eval claims, including Full Duplex Bench gains versus GPT-Realtime-2.1, remain company-reported.
