OpenAI launches GPT-Live voice model; Telnyx ships 16kHz wideband outbound-calling tutorial

OpenAIDevs · x · 2026-09-11

OpenAI's GPT-Live voice model is live, with launch partner Telnyx publishing a full Python tutorial: FastAPI app places outbound calls, streams 16kHz linear PCM bidirectionally (vs. 8kHz G.711) to GPT-Live without resampling, supports tone/emotion/intent nuance, structured tool calls, and session cleanup — a build-ready path for outbound voice agents.

Related event: OpenAI launches full-duplex speech model GPT-Live-1 on API(26 posts)→

Original post →

More from coding & agent

coding & agent channel →