OpenAI Unveils GPT‑Live: Full‑Duplex Voice Redefines ChatGPT Interaction

2 Min Read

Why GPT‑Live’s Full‑Duplex Matters

OpenAI rolled out GPT‑Live today, introducing two models—GPT‑Live‑1 for paid tiers and GPT‑Live‑1 mini for free users—across iOS, Android and the web. Full‑duplex architecture lets the system listen while it speaks, eliminating the silence‑gap that slowed earlier versions. In practice the assistant can interject with acknowledgments like “yeah” or pause for a natural break without cutting you off. The split design separates the voice front‑end from the reasoning engine, which currently runs on the freshly released GPT‑5.5, allowing asynchronous computation while the dialogue continues. 150 million weekly voice users will notice richer visual cards—weather, stocks, maps—appearing inline during conversation. Safety layers have been tightened: synthetic tests show illicit‑behavior scores climbing to ↑0.97 from 0.63, and self‑harm detection reaching ↑0.98. Competitors Google, ByteDance and Nvidia already ship full‑duplex tools, but OpenAI leverages its massive user base and the GPT‑5.5 engine.

“While we release new frontier models, we’ll continuously update the model used by GPT‑Live,” the company said in its blog.

The rollout does not yet support video or screen sharing, and language coverage remains uneven, but the move marks a decisive shift from turn‑based chat toward a more natural spoken interface. Reuters and Bloomberg have noted the strategic importance of voice as the next UI frontier.

Intel provided by Julian Reed (Consumer Electronics Expert).