Published: October 21, 2026
Video Model
Kling 4.0 is the follow-up to Kling 3, one of the most popular video models out there.
The headline change is Middle Frame, length: up to 30 seconds per generation, double Kling 3’s cap, plus 21:9 and stereo audio with multilingual dialogue.
Variants at a Glance
Type | Modalities | Max Res. | Talking Points | |
|---|---|---|---|---|
| Variant | ||||
| Kling 4.0 | Video | T2V, I2V, Omni | 1080p | 30-second single takes, middle frames, stereo audio |
Key Features
Generation
- Image-to-Video with Middle Frames: Start from an image, add up to 8 middle frames and an end frame to steer camera moves, shot changes and reveals.
- Omni (Multi-Reference): Combine up to 15 references in one generation, including video clips as input — up to 10 images and 5 videos.
- Complex Motion: More stable in fights, sports, dance and large-scale action than Kling 3.
Audio & Text
- Stereo Audio: Two-channel sound produced in the same pass, with tighter lip-sync than Kling 3.
- Multilingual Dialogue: 9 languages, with accents and dialects like Cantonese and Indian English.
- On-Screen Text: Renders text in the same languages, plus emoji.
Technical Capabilities
Details | |
|---|---|
| Modalities | Text-to-Video, Image-to-Video (start, middle and end frames), Omni (image and video references) |
| Quality / Resolution | 720p, 1080p |
| Duration | 3–30 seconds |
| Omni Video Input | 3–30 seconds per clip, 30 seconds total |
| Aspect Ratios | Auto, 21:9, 16:9, 9:16, 1:1; image-to-video follows the start frame |
| Audio | Stereo, generated in the same pass; always on |
| Languages | English, Chinese, Japanese, Korean, Spanish, Portuguese, German, French, Hindi |
Limitations
- Up to 1080p in Artlist. Kling’s 4K, 10-bit HDR and video editing aren’t offered here yet.
- Kling 4.0 always generates sound; there’s no audio-off switch. For silent b-roll, strip the track.
Prompting Tips
- Write long clips as timed beats ("0–8s: wide establishing; 8–15s: slow push to close-up").
- Put dialogue and on-screen text in quotes; name the language.