Omni Reference
Combine subject, image and video references to guide appearance, action and camera language. Up to 15 reference items can work together in one generation, helping a character or product stay recognizable across a scene.
Example source
NEXT-GENERATION AI VIDEO
More realism. Greater control. Longer stories. Discover cinematic video with Kling 4.0.
Kling 4.0 is now officially available on this site. Create 3–30 second videos in 720p, 1080p or 4K from text, from images (start and end frames with up to 8 keyframes in between, or up to 10 reference images), or from mixed references that combine images, videos and voice clips.
OFFICIAL MODEL SPECIFICATIONS
CREATIVE POSSIBILITIES
Kling 4.0 is Kling AI’s next-generation video model for text-to-video, image-to-video and reference-guided creation. Its official announcement brings together longer scenes, richer audiovisual detail and more precise control over characters, camera movement and story beats.
Combine subject, image and video references to guide appearance, action and camera language. Up to 15 reference items can work together in one generation, helping a character or product stay recognizable across a scene.
Example sourcePlan up to 10 keyframe images to define important visual moments. Guide a scene from its opening through a turn in the action to its final frame, with single-generation video lengths of 3–30 seconds.
Example sourceDescribe a tracking shot, push-in, orbit or continuous action. Kling 4.0 focuses on clearer motion and smoother shot transitions, making it useful for action scenes, product reveals and film previsualization.
Two-channel stereo audio and improved lip-sync connect dialogue, ambient sound and character performance. The announcement lists Chinese, English, Japanese, Korean, Spanish, Portuguese, German, French and Hindi, among other languages.
Output options include 720p, 1080p and 4K, with 21:9, 16:9, 1:1 and 9:16 aspect ratios. More natural facial expressions, lighting and textures help support close-ups and cinematic composition.
Reference-guided editing targets subjects, backgrounds, movement or visual style. Input editing supports up to five clips, including one primary video, with a combined length no longer than 30 seconds.
Source: Kling AI release notes These are the officially announced model capabilities. On this site, Kling 4.0 supports text-to-video, image-to-video and mixed-reference video in 720p, 1080p and 4K.
SEE THE IDEAS IN MOTION
Watch three examples featured by the reference site. Look at shot continuity, camera movement and subtle character performance before planning your own scene.
A wide ocean scene, the performer and close-ups of hands form a continuous musical story.
Example sourceA camera follows two martial artists through a training-hall exchange, keeping their movement readable.
Example sourceNatural lighting and quiet expressions carry the emotion of a conversation between two characters.
Example sourceTWO MODELS. DIFFERENT NEEDS.
| Specification | Kling 4.0 | Kling 4.0 Flash |
|---|---|---|
| Single-generation length | 3–30 seconds | 3–20 seconds |
| Listed resolution | 720p / 1080p / 4K | 720p |
| Image input | Image-to-video, first & last frames | Image-to-video from a first frame |
| Multi-keyframes | Up to 10 keyframe images | Not listed in the Flash specification |
| Omni Reference | Listed | Listed |
This comparison follows the official announcement and is not a quality ranking.Source: Kling AI release notes
FROM IDEA TO CREATIVE BRIEF
A focused prompt and a small set of useful references are a good starting point for both text-to-video and image-to-video. Prepare the direction now, then select an available model in the studio.
Name the subject, location and one main action. Choose the feeling and the most important moment rather than asking for several unrelated scenes.
Use a character image, product photo or movement reference where it adds clarity. Decide what must stay consistent and which moments need keyframes.
Open the studio and check the model, duration, ratio and resolution. Kling 4.0 is available here for 3–30 second videos in 720p, 1080p and 4K.
Watch for identity changes, unexpected cuts and mismatched motion. Change one instruction at a time so you can see which adjustment improved the result.
Use these original practice prompts as creative starting points. They are briefs, not claims about generated results.
A pianist performs on a ship at sunset. Begin with a wide view of the ocean, track slowly toward the piano, then finish on the hands. Keep the performer and piano consistent. Natural wind, moving waves and warm cinematic light.
Use this prompt in the studioA ceramic perfume bottle rests on wet stone. A slow camera orbit reveals its shape, followed by a close-up of the surface and label. Soft side lighting, subtle water reflections, one continuous movement, no changes to the product design.
Use this prompt in the studioTwo old friends recognize each other on a quiet street at dusk. Follow their approach in one continuous shot, then hold on the smile before they embrace. Consistent clothing and faces, natural footsteps, gentle handheld camera movement.
Use this prompt in the studioKling 4.0 is available in the studio. Check the credit estimate for your settings before generating. Our pricing page explains the current platform plans.
Kling 4.0, also searched as Kling4 or Kling AI 4.0, is the next-generation Kling AI video model described in the official release note. It brings text-to-video, image-to-video, richer references, multi-keyframe control and synchronized audio into a more complete video workflow.
Kling 4.0 has officially launched and is available on this site now. Select Kling 4.0 in the video studio to start creating.
Yes. Kling 4.0 is now officially available on this site. Create 3–30 second videos in 720p, 1080p or 4K from text, from images (start and end frames with up to 8 keyframes in between, or up to 10 reference images), or from mixed references that combine images, videos and voice clips.
Yes. Kling 4.0 on this site supports 4K output as well as 720p and 1080p. Choose the resolution in the studio before generating.
Each Kling 4.0 video on this site can be 3–30 seconds long, set in one-second steps.
In image-to-video you can use a start frame, an end frame and up to 8 keyframes in between, for up to 10 frames in total (longer videos allow more frames; all 10 are available from 9 seconds), or up to 10 reference images. With mixed references you can combine up to 10 images, 5 videos (30 seconds in total) and 3 voice clips, with up to 15 files in one generation.
The standard model lists longer single-generation clips, 4K and multi-keyframes. Flash is positioned for faster iteration, with 3–20-second clips, 720p and Omni Reference. The Flash specification does not list multi-keyframes, so this page does not assume they are supported.
Yes. Kling 4.0 generates video with synchronized sound, including dialogue and lip-sync. With mixed references you can also upload up to 3 voice clips (5–15 seconds each, used together with an image or video) to guide character voices, or keep the original sound of a reference video.
You can watch the video examples on this page for free. Generation uses the models available in the studio and your credit balance; free credits depend on account offers. Check the selected model’s credit cost and platform plans before generating.
A Kling 4.0 API is not offered by this page. Consult the platform API documentation for supported models and endpoints. Official Kling AI access and this website’s integrations are separate.
Bring a prompt, an image or a reference. Explore the available Kling workflow and give your next scene a direction.
Start creating