
Tell a short story in one scene
A complete beat in one generation: setup, action, and payoff. Up to 30s, with multishot and timed prompts.
Tell your story
Kling 4.0 on Magnific
Kling 4.0 is Kuaishou's AI video model for longer scenes.
Longer AI video scenes, up to 30s in 1080p, with native audio and keyframes to direct every move.
For the final cut. Scenes from 3–30s in 720p or 1080p, with native stereo audio, up to 10 keyframes, and up to 15 mixed references.
For exploring ideas fast. Scenes from 3–20s in 720p.
Launch film
The launch film for Kling 4.0, from Magnific.
What's new
Up to 30s in a single generation, so the action has time to land. Pick anything from 3–30s, or keep the 10s default.
Omni Reference mixes image, video, subject, and voice inputs, with up to 15 references in total.
Best for
Four ways to start, from a single scene to a brand film.

A complete beat in one generation: setup, action, and payoff. Up to 30s, with multishot and timed prompts.
Tell your story
Start from a designed image and decide how it moves, with first- and last-frame control and up to 10 keyframes.
Animate a still
Dialogue with lip sync in 9 languages and several accents, all in the same generation.
Generate a talking scene
Keep characters, products, and style consistent across scenes, with on-screen text, using Omni Reference.
Build a brand filmSpecs
Everything you need to know before your first scene.
| Inputs | Text, Image, Video |
|---|---|
| Max length | 3–30sKling 4.0 Flash: 3–20s |
| Resolution | 720p, 1080p |
| Frame rate | — |
| Native audio | YesStereo, with lip-sync in 9 languages |
| References | Up to 15Including video |
| Aspect ratios | 21:9, 16:9, 1:1, 9:16, auto |
| Credits / s | — |
| Plans | Credits on any paid plan |
Credits per second, plans and some Flash values are still to be confirmed on Magnific.
Credits per second at list price. Where a line shows 2 figures, the note under it says what each one is.
Compare every Kling versionKling 4.0 is Kuaishou's AI video model for longer scenes. On Magnific, it generates 3 to 30 seconds in a single generation, at 720p or 1080p, with native stereo audio, up to 10 keyframes, and up to 15 mixed references.
From 3 to 30 seconds in a single generation, with 10 seconds by default.
You can choose 720p or 1080p.
Yes. It generates native stereo audio with lip sync, and voice in 9 languages and several accents. Audio is always on.
Up to 15 in total, mixing image, video, subject, and voice inputs. Each type has its own limit: 10 images, 5 videos (30 seconds combined), 7 subjects, and 1 voice reference. Together, they can't go over 15.

Compare every Kling version and pick the one that fits the shot.