# ElevenLabs launches Scribe v3 and Scribe v3 Live speech-to-text models

> Original title: Introducing Scribe v3 and v3 Live, our most accurate speech to text models

- Source: ElevenLabs Blog
- Published: 2026-11-10T12:00:00.000Z
- Bad Influence: https://badinfluence.news/items/a0dmgbbxbddf74r9rvw7rit2y
- Original: https://elevenlabs.io/blog/scribe-v3

## Summary

ElevenLabs launches Scribe v3 for recorded audio, which makes 27% fewer word errors than Scribe v2 across 102 languages. Scribe v3 Live, its realtime variant, makes 43% fewer errors than Scribe v2 Realtime and adds speaker diarization. Both models are available in ElevenAPI, and switching from Scribe v2 requires changing only the model ID.

## Why it matters

The source gives measured error reductions against Scribe v2 across 102 languages and explains which model fits recorded audio versus realtime use.