All models

ElevenLabs AI models

6 models · from 1 credits

At a glance

Models
6
Lowest starting price
ElevenLabs Scribe V3 · 1 credit
Newest
ElevenLabs Scribe V3 · November 10, 2026
Types
Audio

The models

Prices in US dollars at list price ($1 = 1,000 credits).
  • ElevenLabs Scribe V3ElevenLabs · Audio · per second

    New

    Word-level transcripts with speaker labels and sound events in 102 languages

    • 112 languages

    $0.0002

    Try
  • ElevenLabs v4 TurboElevenLabs · Audio · per 1,000 characters

    Faster ElevenLabs v4 speech with the same 21 voices and 90+ languages

    • 21 voices
    • 91 languages

    $0.048

    Try
  • ElevenLabs v4ElevenLabs · Audio · per 1,000 characters

    ElevenLabs v4 speech in 21 voices and 90+ languages, up to 10,000 characters

    • 21 voices
    • 91 languages

    $0.096

    Try
  • ElevenLabs Voice ChangerElevenLabs · Audio · per second

    Re-voice a recording in one of 41 preset voices, same words and timing

    • 41 voices

    $0.005

    Try
  • ElevenLabs DubbingElevenLabs · Audio · per second

    Dub a recording or a video's soundtrack into 33 languages, up to 15 minutes

    $0.012

    Try

Choosing an ElevenLabs model for voiceovers, dubbing and clean speech

Five tools for the voice in your project

The ElevenLabs group holds five models, and each one starts from something different. ElevenLabs v4 and v4 Turbo read a script you type. Dubbing translates a recording or a video's soundtrack into another language. Voice Changer keeps your own performance and gives it a different voice. Audio Isolation pulls clear speech out of a noisy recording. Pick by what you have in hand: words on a page, a recording in the wrong language, a performance in the wrong voice, or a take full of background noise.

Reading a script with v4 or v4 Turbo

Both voice models take up to 10,000 characters at a time, offer the same 21 voices and cover more than 90 languages, so a script can move between them without changes. Turbo costs half as much per character, which makes it the easy choice for drafts, tutorials and lots of short lines; v4 is the one to keep for the final read. Write cues such as [whispers] or [laughs] in square brackets where they should happen, lower stability for a livelier delivery, and set the language when names or mixed languages trip up the accent.

Working from a recording

Dubbing takes a recording or a video up to 15 minutes and returns it in one of 33 languages, keeping the background sound unless you turn that off. For a video, only the soundtrack changes: the picture stays as filmed and the lips are not redrawn, so pair it with a lip sync tool when the speaker fills the frame. Voice Changer takes up to 5 minutes and swaps the voice for one of 41 presets while your pauses and emphasis stay. Audio Isolation keeps the speech and drops fans, traffic and music from recordings up to an hour. Use only your own voice or recordings you have permission to use.

Quick answers

Which ElevenLabs model should I use for a voiceover?

Start with ElevenLabs v4 Turbo while you work on the wording, since it costs half as much per character, then render the final take with ElevenLabs v4. Both use the same 21 voices and settings, so the voice you pick for the draft is available for the final read.

Can ElevenLabs translate a video?

Yes. ElevenLabs Dubbing turns a video's speech into one of 33 languages and returns the same picture with a new soundtrack, for videos up to 15 minutes. It does not move the lips or add subtitles, so use a lip sync tool afterwards when the face is on screen.

What is the difference between Voice Changer and text to speech?

Text to speech reads words you type in a voice you choose. Voice Changer starts from your own recording and keeps its timing, pauses and emphasis, only swapping the voice. Choose Voice Changer when the delivery matters, such as acting a character or matching a scene.

How do I clean up a noisy recording?

Use ElevenLabs Audio Isolation. Upload the recording or a video as it is, with no settings to choose, and you get an MP3 of the speech with background noise, music and room sound taken out. It is made for speech; to split a song into vocals and music, use a vocal remover instead.

Do I need permission to change or translate someone's voice?

Yes. Voice Changer and Dubbing work on a real person's voice, so only upload your own recordings or ones where the speaker agreed. PopNow asks you to confirm this the first time you use these tools, and the same rule applies every time after that.