OpenAI Updates GPT Live 1.1 Voice Mode: 13 New Voices and Near-Zero LatencyBrief · digest for October 20
Brief≈ 1 min

OpenAI Updates GPT Live 1.1 Voice Mode: 13 New Voices and Near-Zero Latency

If you've ever used voice chat with an AI, you know that awkward walkie-talkie rhythm: you dictate your prompt, wait for transcription, and wait again for speech synthesis. With OpenAI's GPT Live 1.1 update, that lag is finally gone.

The model now processes audio tokens end-to-end (full-duplex) without converting speech to text in the middle. Response latency drops to a minimum: the conversation flows in real time, and you can interrupt the assistant mid-sentence—the model stops immediately and pivots to your new thought.

Key details from the release:

  • 13 new voices across 6 languages: natural inflections for French, German, Hindi, Mexican Spanish, Korean, Japanese, plus an expressive English voice called Bobby.
  • Cleaner conversation: filler phrases like "checking..." during web search calls have been stripped out.
  • Broad availability: live across the ChatGPT mobile apps, web, Codex, and the Realtime API.

For entry-level plans, OpenAI also released a lighter GPT-Live-1.1 Quip variant. Earlier, we looked at how real-time voice assistants evolved and how voice-driven feature development works with Codex—and now that dialogue actually feels alive.

Want a deep diveVotes help us choose what to break down in detail. This material has 0 votes.
Sources · 2