Prompt
Text · Required
Loading the current model…
We could not open this model.
Return to the catalogue and choose an available model.
Back to the catalogue
- Kuaishou
- Standard
- Generate video
Kling 3 Omni
Combine multimodal references into controlled video, motion, and sound.
Compare Kling 3 Omni settings and cost estimates. Reference-rich Kling projects where the catalogue's image and video reference inputs are useful for describing the intended scene.
- Generate video
Kling 3 Omni pricing example
$0.2822 Estimated per run
- Aspect ratio: 16:9
- Duration: 3
- With audio: false
Estimate checked . This example uses the configuration shown above. Inputs and settings can change the price; check the estimate for your own run before generating. Project storage is billed separately.
How AI usage and project pricing work →Curated output references.
What goes in. What comes back.
The model defines every slot it reads and every result it returns. Required inputs must be present before a run can begin.
Prompts and source material
Negative prompt
Text · Optional
First frame
Image and Video · Optional
Last frame
Image and Video · Optional
Image reference
Image · Optional · Up to 5
Video reference
Video · Optional
Files and structured results
Video
Video
Configure the run.
| Control | Available values | Requirement |
|---|---|---|
| Aspect ratio | 16:9, 9:16, and 1:1 | Optional |
| Duration | 3–15 | Optional |
| With audio | On, Off | Optional |
Works especially well for
- Reference-rich Kling projects where the catalogue's image and video reference inputs are useful for describing the intended scene.
Know the trade-offs
- The Omni catalogue label does not guarantee all vendor Omni editing modes. Available controls are 3–15 seconds, audio, first frame, and reference inputs; validate a short shot first.
Continue from the result
Use it inside a production
- Suggested exercise: use a product reference and a room reference to direct a restrained reveal, checking that the product keeps its identity as the framing changes. →
- Choose Kling 3 for a straightforward text/first-frame shot; compare Seedance 2.0 for another workflow with multiple image and video references. →
Compare similar models
Compare models for the same capability. Check their inputs, output settings, and estimate using the same brief.
Configure the model.
Keep the result in the cut.
The generation returns to the scene, asset folder and timeline it belongs to.