Storytellers and video creators
Develop short narratives with planned story beats, consistent characters and dialogue. Use a longer take when the scene needs time to build.
Kling 4.0 generates videos with synchronized stereo sound from text, images and mixed references. Plan scenes up to 30 seconds long, guide key visual moments, and make targeted changes to existing footage.
Explore six Kling AI examples, then use a starting prompt to create your own version. Starting prompts describe the scenes; they do not reproduce the original reference or editing workflow.
Kling 4.0
A Lion Dance from Storyboard Panels
Kling 4.0
A Clay Astronaut Dog on a Cheese Moon
Kling 4.0
A Collage Poster in Motion
Kling 4.0
A Fantasy Scene Guided by Sketches
Kling 4.0
A New Character Following a Reference Camera Move
Kling 4.0
A Street Performance with Passers-by Removed
A multimodal video model that combines reference guidance, synchronized sound and editing.
Kling 4.0 is a video model developed by Kuaishou and released by Kling AI on October 15, 2026. It brings video generation and editing into one model, with longer scenes, richer reference inputs and control over important moments in a story.
Images can guide a character, object, composition or style. Video references can guide movement and camera work, while reusable elements and voice profiles help define recurring subjects.
Start with the material that best describes your idea:
The specifications below describe the official Kling 4.0 model. Available controls and output options depend on the creation interface.
Control story length, references, timing, sound, movement and edits within one video model.
Generate 3–30 seconds in one pass, giving an action, conversation or short story time to unfold. Use a continuous take or describe a sequence of shots.
Guide appearance, composition, motion and vocal identity with up to 15 mixed assets. Separate limits apply to images, videos, elements and voice profiles.
Use keyframe images to guide character states, scene changes and reveals. A first and last frame can also define the beginning and ending of an image-to-video clip.
Generate dialogue, singing, ambience and sound effects with two-channel stereo audio. Lip synchronization helps match the voice to the character's performance.
Use reference footage to guide action, pacing and camera movement while changing the subject or setting. Prompts can also describe tracking shots, pans and orbits.
Replace a character or background, remove an unwanted element, or change a visual style. Describe both the intended change and the scene details you want to preserve.
Bring personal stories, animation, visual communication and video revisions into motion.
Develop short narratives with planned story beats, consistent characters and dialogue. Use a longer take when the scene needs time to build.
Turn character sheets, drawings and storyboards into moving scenes. Guide the visual style with references, from clay animation to illustrated fantasy.
Animate posters, demonstrate an idea or create a product sequence from visual references. Combine motion and sound to make the message easier to follow.
Revise existing footage with targeted instructions, such as removing a passer-by or replacing a setting. Keep the unchanged parts of the shot clearly specified.
Build a clear brief, direct the sequence and review the result.
Enter a text prompt, upload a starting image, arrange keyframes or add image and video references. Choose the input mode that fits your material.
Describe the subject, action, camera and sound. Select 720p or 1080p, a duration from 3 to 30 seconds and an available aspect ratio, then review the displayed credit cost.
Generate the clip and review motion, consistency, dialogue and visible text. Refine the prompt and generate another version before downloading your chosen result.
Official model specifications; individual creation interfaces may expose a subset of these options.
| Developer | Kuaishou / Kling AI |
| Release date | October 15, 2026 |
| Modes | Text to video; image to video; keyframe to video; reference to video; video editing |
| Duration | 3–30 seconds; 1-second increments |
| Output resolutions | 720p and 1080p |
| Aspect ratios | 21:9, 16:9, 1:1, 9:16; Auto in Omni Reference |
| Audio | Two-channel stereo with synchronized speech, singing and sound effects |
| Keyframes | Up to 10 keyframe images |
| Mixed references | Up to 15 assets total: up to 10 images, 5 videos, 7 elements and 3 voice profiles, within the combined cap |
| Shared video quota | Up to 5 clips and 30 seconds combined across reference videos, editing videos and video elements |
| Text input | Up to 8,000 tokens |
| Editing source | One primary video; each input clip must be 3–30 seconds and fit the shared video quota |
Last updated: November 10, 2026
Official model-level comparison. The Kling 3.0 series includes Omni workflows; these are not the settings of a single generator.
| Spec | Kling 4.0 | Kling 3.0 Series |
|---|---|---|
| Native clip duration | 3–30 seconds | 3–15 seconds |
| Text and image to video | Supported | Supported |
| First and last frames | Supported | Supported |
| Multiple keyframes | Up to 10 | Not supported in the official comparison |
| Audio channels | Two-channel stereo | Single-channel |
| Reference videos | Up to 5; 30 seconds combined | Up to 1; 15 seconds maximum |
| Combined references | Up to 15 mixed assets | Up to 7 images/subjects, or 1 video with 4 images/subjects |
| Video editing | Integrated multimodal editing | Available through 3.0 Omni workflows |
Kling 4.0 Advantages: More room for extended scenes, additional reference assets, multiple keyframes and stereo sound.
Kling 3.0 Series Advantages: A focused fit for shorter scenes that need text or image guidance, start/end frames and native audio.
Choose Kling 4.0 when a story needs longer timing, richer references or finer visual control. Kling 3.0 suits projects whose scene and reference requirements fit within its shorter generation range.
Compare Shotyard subscription plans and their included credits. Model-specific generation costs are shown in the generator before submission.
Compare output-only credit costs for common resolution and duration settings. The generator calculates the total for your selected inputs.
| Resolution / duration | Credits per generation |
|---|---|
| 720p / 5 seconds | 150 |
| 1080p / 5 seconds | 200 |
| 720p / 10 seconds | 300 |
| 1080p / 10 seconds | 400 |
For reference-to-video, if the combined reference-video duration exceeds the output duration, half the excess is added to the billable seconds. Other modes have no reference-video surcharge.
Answers about inputs, timing, audio, editing and how to guide your results.
Kling 4.0 is Kuaishou's multimodal video model, released by Kling AI on October 15, 2026. It generates videos with audio from text and visual references, and also supports editing existing clips.
Native generation runs from 3 to 30 seconds, in 1-second increments. A 30-second generation can be a continuous take or a sequence with several story beats.
Yes. It generates two-channel stereo audio, including dialogue, singing, ambience and sound effects. Describe the sound you want alongside the visual action, and review lip synchronization in the result.
Omni Reference accepts up to 15 assets in total, with separate limits of 10 images, 5 videos, 7 elements and 3 voice profiles. These are individual ceilings, not quantities you can all add together. Reference videos, editing videos and video elements also share a maximum of 5 clips and 30 seconds combined.
Keyframes define visual moments that the sequence should pass through, with up to 10 images. References describe things such as a character's appearance, scene layout, style or camera movement without necessarily fixing a particular frame.
Yes. Instructions can guide changes to characters, backgrounds, movement, camera direction or style, and can remove selected elements. Identify one primary video and state which parts of the scene should stay the same.
Create at 720p or 1080p with 21:9, 16:9, 1:1 or 9:16 framing. Auto aspect ratio is also available in the generator.
Describe the subject, setting, action, camera and sound. When using references, assign each one a role: for example, use @Image1 for appearance and @Video1 for camera movement. For edits, explain the change and the details to preserve.
Shotyard offers plans with included credits. Review the plan terms and the generation cost displayed for your selected settings before submitting a task.
You can use the Kling 4.0 AI Video Generator on Shotyard.