Fish Audio launches Live 3: a full-duplex voice model that listens while talking, sub-200ms reactions
EXM7777 · x · 2026-11-11
Fish Audio introduced Live 3, a full-duplex voice model that keeps listening while it speaks, says "mm-hmm," and knows when to take turns. Key details:
- One full-duplex model replaces the chain of separate speech, LLM, and voice models; audio streams in and out simultaneously
- It reads where you are in a sentence from the audio itself, not silence timers
- Turn-taking learned from real conversations, not hardcoded rules
- Quick reactions fire in under 200ms; a separate "thinker" model drafts full answers in the background
The author argues this makes AI call centers, receptionists, and sales agents "10x more human overnight." Fish Audio is offering free custom agents for the first 1,000 submitted use cases.
Related event: Fish Audio launches full-duplex voice model Live 3(8 posts)→
More from Multimodal
- Japanese user's Fable 5.2 showcase video goes viral across languages — FinanceYF5 · 2026-11-11
- One-sentence prompt to Fable 5.2 yields jaw-dropping self-showcase video — FinanceYF5 · 2026-11-11
- AI-made music video "Slopcore, Encore" enters Claudia contest via Claude skill + Fable 5.2 — AIandDesign · 2026-11-11
- TRIAD benchmark: no multimodal retrieval system leads across its six subsets — _reachsumit · 2026-11-11
- Vision tokens far denser than text: paper traces VLM modality gap to early-layer processing — _reachsumit · 2026-11-11
- Fable 5.2 one-prompt short film "936": "make a video that will deeply change my life" — yuwen_lu_ · 2026-11-11