ElevenLabs ships Scribe v3 and Scribe v3 Live speech-to-text models with lower error rates
ElevenLabs released Scribe v3 for recorded audio and Scribe v3 Live for realtime transcription, cutting word errors by 27% and semantic errors by 43% across 102 languages, with diarization in streaming and 93 ms time-to-first-sound.
Read the full story at AlphaSignal ↗
Timeline · 8 reports
- 2026-11-12 18:00 · DEV Community — Machine Learning
ElevenLabs launches Scribe v3 and Scribe v3 Live, its most accurate speech-to-text models yet - 2026-11-12 13:48 · The Decoder
ElevenLabs Scribe v3 Live uses conversational context to stop voice agents from cutting off callers mid-sentence - 2026-11-12 07:20 · DEV Community — AI
ElevenLabs Scribe v3 Live Cuts Semantic Errors 43% — Dual-Track WASAPI Still Sets the Latency Floor - 2026-11-12 04:14 · DEV Community — AI
Fish Audio Live 3 Claims 186 ms Full-Duplex Reactions — Dual-Track WASAPI Still Sets the Real Floor - 2026-11-11 14:23 · DEV Community — AI
ElevenLabs Scribe v3 Live Lands at 93 ms TTFS — Dual-Track WASAPI Capture Still Sets the Overlay Floor - 2026-11-11 07:19 · DEV Community — AI
Claude Meetings Joins the Call — Your Overlay Still Lives on Dual-Track WASAPI Capture - 2026-11-11 04:43 · MarkTechPost
Fish Audio Releases Live 3, a Full-Duplex Voice Model That Reacts in 186 ms - 2026-11-10 18:16 · AlphaSignal
ElevenLabs' Scribe v3 Cuts Live Transcription Errors by 43% Across 102 Languages