OpenAI · 2026-11-19 · notable
GPT-Live-1.1 — OpenAI's voice model gets steadier multilingual accents
In short
GPT-Live-1.1 is OpenAI's new full-duplex voice model in the API, with better pronunciation, accent consistency and transcription across languages. A new standby mode bills idle WebRTC calls at $0.02 per minute.
OpenAI · 2026-11-19 · 1 source checked

OpenAI's second Live voice model, plus a cheap standby mode for quiet calls.
What is it?
GPT-Live-1.1 arrived in the OpenAI API on November 19, 2026 as the model id `gpt-live-1.1`. OpenAI's changelog says it improves pronunciation, accent consistency and transcription in multilingual voice conversations compared with GPT-Live-1, and it works with existing Live session configurations.
How does it work?
The model is full-duplex: it takes audio and text in, returns audio and text, and keeps listening while it speaks. It runs only on the Live endpoint (`v1/live/sessions`) with streaming and function calling. Standby, added the same day for GPT-Live-1 and 1.1 WebRTC sessions, pauses inference after an idle timeout of 5 to 60 seconds and resumes when the user speaks, keeping the same session.
Why does it matter?
Voice agents that switch languages or read back names and numbers get a more consistent voice without code changes. Long calls with quiet stretches get cheaper: active time costs $0.05 per minute and standby $0.02, with backend model and tool use billed separately.
Who is it for?
developers building voice agents
Try it
Set the model to gpt-live-1.1 in a Live session (v1/live/sessions)