Scribe v2
ElevenLabs’ highly accurate transcription model.
Scribe v2 is ElevenLabs’ speech-to-text model, consistently among the most accurate on Artificial Analysis’ AA-WER benchmark.
- Provider
- ElevenLabs
- Type
- Speech-to-text
- Released
- Jan 9, 2026
- Price
- ≈$3.67 per 1,000 minutes
- Input
- audio
- Output
- text
- Open weights
- No
Best for
- Extraction & classification
- Multimodal
Strengths
- ≈2.2% WER
- Speaker diarization
More from ElevenLabs
- Eleven v4 — ElevenLabs. The most preferred voice in the Artificial Analysis speech arena.