> ## Documentation Index
> Fetch the complete documentation index at: https://elarislabs-87.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Video models

> Six major video generation families — Google Veo, Kling, Sora, LTX, Seedance, and utility models — covering T2V, I2V, lip sync, and upscaling, all priced per second of output.

Video credits scale with **output duration × model rate**, plus resolution and audio multipliers. Studio displays the estimated cost before you confirm every generation.

> [!TIP]
>
> MCP video tools accept `durationSeconds` (typically 2–12 seconds). Studio shows the cost estimate before you run — adjust duration or model tier to control spend.

## Which video model should I use?

> [tab] Google Veo
>
> **High-quality text-to-video and image-to-video with built-in audio support.**
>
> | Model | Notes |
> | - | - |
> | Veo 3.1 | Flagship; full audio support |
> | Veo 3.1 Fast | Faster generation at the same quality tier |
> | Veo 3.1 Lite | Efficient variant; lower per-second cost |
> | Veo 3 | Previous generation; solid quality |
>
> Best for: cinematic scenes, complex motion, and productions that need built-in audio synced to video.

> [tab] Kling
>
> **Versatile motion and general video generation.**
>
> | Model | Notes |
> | - | - |
> | Kling O3 | Latest generation; strong motion control |
> | Kling V3 | General-purpose video |
>
> Best for: character animation, motion-heavy clips, and mid-tier budget productions.

> [tab] OpenAI / LTX / ByteDance
>
> Three distinct families covering cinematic, efficient, and social-ready outputs.
>
> | Model | Provider | Notes |
> | - | - | - |
> | Sora 2 | OpenAI | Cinematic quality; strong scene coherence |
> | LTX 2.3 | Lightricks | Efficient generation; good for fast iteration |
> | Seedance 2.0 | ByteDance | Social-ready motion; fast turnaround |
> | Seedance 2.0 Fast | ByteDance | Fastest variant; lower cost per second |

> [tab] Utilities & Specialized
>
> Specialized models for editing, transformation, and lip sync.
>
> | Model | Notes |
> | - | - |
> | Happy Horse | Distinctive motion style |
> | Gemini Omni Flash | Multimodal video generation tasks |
> | Fabric | Lip sync — matches audio track to existing talking-head video |
>
> > [!NOTE]
> >
> > Utility transformations (inpaint, background removal, upscale, extend) are available for post-processing existing clips and have fixed-ish action costs rather than per-second pricing.

## What capabilities are supported?

> [card] Generation
>
> Text-to-video, image-to-video, reference-to-video, and first / last frame anchoring

> [card] Motion control
>
> Camera motion, motion intensity controls, and motion reference inputs

> [card] Lip sync
>
> Sync dialogue audio to an existing talking-head video (Fabric model)

> [card] Post-processing
>
> Inpaint, background removal, upscale, and extend existing clips

> [!TIP]
>
> Use [HyperFrames](https://elarislabs-87.mintlify.site/studio/hyperframes.md) when you need a multi-stage animated pipeline — compositing multiple clips, adding transitions, and layering effects — rather than a single generated clip.

> [card] [How pricing works](https://elarislabs-87.mintlify.site/pricing/how-pricing-works.md)
>
> Full breakdown of per-second rates, duration, and resolution multipliers.

> [card] [Image models](https://elarislabs-87.mintlify.site/models/image.md)
>
> Browse the image model catalog with per-generation credit costs.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.