Video · Character.AI

Ovi

An open-source model that animates an image into a 5 s clip with speech and sound, in private mode. Pick it for a cheap, private talking shot.

Private Low cost
Latest versionOviOctober 5, 2025
Price$0.0198 per second of video$0.099 a 5 s clip
Key specClips up to 5 s, with sound

What sets it apart

Ovi from Character.AI was billed as the first open-source video model with native audio, with 11B parameters and an Apache 2.0 license. It makes short 5 s clips with speech, effects and ambience from one image and a prompt.

  • Open weights, Apache 2.0
  • Tagged dialogue and sound prompts
  • 5 s clips with audio

Strengths and limits

Where it shines

  • Speech and sound generated with the picture
  • Among the cheapest clips with sound
  • Private mode

Think twice when

  • 5 s only, from an image
  • Lower resolution than recent commercial models
  • 5 s, native sound
  • Image to video only

Use cases

Talking portrait

Animate this portrait: <S>Welcome to our shop, take your time.<E> <AUDCAP>Soft jazz, quiet street sounds<ENDAUDCAP>

Tags place speech and ambient sound precisely.

Two-voice exchange

Two travelers at a train platform. <S>Is this the right train?<E> <S>Yes, platform four.<E> <AUDCAP>Station announcements, rolling suitcases<ENDAUDCAP>

It handles multi-speaker dialogue.

Cheap sound test

A rain-soaked window with a cat watching droplets, <AUDCAP>steady rain, distant thunder<ENDAUDCAP>

Low price per second with native audio.

Against its rivals

  • LTX VideoLTX is larger with 4K, multi-shot and longer clips.#15 of 22 on Artificial Analysis Image to Video Arena (1,038 Elo)
  • LongCat VideoLongCat makes 30 s silent clips.
  • Wan 2.xWan 2.x makes 5 to 15 s clips up to 1080p.#8 of 22 on Artificial Analysis Image to Video Arena (1,077 Elo)
Compare side by side ↗

Version history

Each version, dated, with what it changed over the one before.

  1. OviLatest
    • First open-source video model with native audio, from text or text plus image.
    • Multi-speaker dialogue with timing controlled by tags in the prompt.
    Source : blog.fal.ai
  2. Ovi 1.1Earlier, not in Sub Rosa
    • Adds 10 s videos at 960x960 resolution.
    Source : github.com

The model in this family

ModelUsePrivacyPrice
Oviovi-image-to-videoImage to videoPrivate$0.0198 per second of video$0.099 a 5 s clip

Specifications

SpecificationOvi
ReleasedOctober 5, 2025
Parameters11B
WeightsOpen, Apache 2.0
Acceptstext, images
Producesvideo, audio
Clip length5 s, with sound
Price$0.0198 per second of video
PrivacyPrivate

Prices, context and lengths from the live catalog on October 16, 2026; parameters and licenses from the maker.

Sources

Everything on this page comes from these pages, read on October 16, 2026 or as dated on each score.

Show the 2 sources
Learn more about Ovi ↗