OpenAI launches GPT-Live, a full-duplex voice model replacing Advanced Voice Mode in ChatGPT

OpenAI introduced GPT-Live, a full-duplex voice model family (GPT-Live-1 and GPT-Live-1 mini) that can listen and speak simultaneously, replacing the turn-based Advanced Voice Mode in ChatGPT
The model delegates complex reasoning or search tasks to GPT-5.5 in the background while keeping the conversation flowing, and OpenAI says it will keep updating the underlying frontier model
OpenAI's evaluations show GPT-Live-1 preferred over Advanced Voice Mode in 75.7% of head-to-head conversations, with higher ratings for flow and pleasantness
Hacker News commenters were largely positive about reduced interruption issues and background task delegation, but flagged persistent problems like over-interrupting, jarring filler-word timing, and weak real-time translation quality
OpenAI staff confirmed in comments that GPT-Live currently lacks video support (unlike Advanced Voice Mode) and offers nine selectable voices with imperfect multi-speaker detection