ElevenLabs · audio
ElevenLabs v3
The most expressive ElevenLabs voice model, with audio tags for emotion and delivery.
Text to speech
About ElevenLabs v3
ElevenLabs v3 is the most expressive voice model from ElevenLabs, priced at $0.050 per 1K characters. It runs as text to speech: you pick a voice, Rachel by default, and paste the script.
The price is counted on the characters of the text, so the longest script a request takes costs $1.25. A typical generation takes about 15 s.
Good for
- Voiceovers that need emotion and delivery, steered by audio tags in the text
- Characters and ads where a flat read would not do
Getting good results
- Add audio tags to the script where you want a change of emotion or delivery; the model takes them as directions for the read.
- You pay for the characters you send: $0.050 for a short line, $0.25 for a paragraph.
ElevenLabs v3 questions
ElevenLabs v3 or ElevenLabs Multilingual v2?
They share the same voices; ElevenLabs v3 costs $0.050 per 1K characters and is the expressive model with audio tags; ElevenLabs Multilingual v2 is the steadier one for long narration, at $0.050 per 1K characters.