VQV Signal
According to a post on X by @Marktechpost, Fish Audio has launched Live 3, a voice model that handles turn-taking inside the speech model itself. The model was trained on real conversations to decide when to acknowledge, wait, respond or stop, and Fish Audio says it can react in as little as 186 ms. At launch it supports English only, no benchmarks have been published, and agents cost $0.035 a minute through Dec 31.
According to a post on X by @Marktechpost, Fish Audio has launched Live 3, a voice model that handles turn-taking inside the speech model itself. The model was trained on real conversations to decide when to acknowledge, wait, respond or stop, and Fish Audio says it can react in as little as 186 ms. At launch it supports English only, no benchmarks have been published, and agents cost $0.035 a minute through Dec 31.
AI-assisted summary based on the listed source.
Live 3 handles turn-taking inside the speech model, trained on real conversations to decide when to acknowledge, wait, respond or stop, and Fish Audio says it can react in as little as 186 ms. It is English only at launch with no published benchmarks; agents cost $0.035 a minute through Dec 31. Why it matters:...
Teams building AI phone support and other voice agents would be most affected, since the claimed ability to stop talking over callers addresses a common usability problem. Without published benchmarks, the latency and turn-taking claims cannot yet be compared with other systems. Watch for independent testing, language support beyond English, and pricing after Dec 31.
VQV organizes public signals from inspectable sources. It does not independently verify the underlying report.
Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.
VQV surfaced this signal because it is recent, relevant to AI Voice, connected to Scobleizer News Pulse (X).
No login, cookies, social SDKs, or automatic posting.
