Skip to content

Kling 4.0: plan the scene, choose the version

Kling 4.0 is Kuaishou’s announced video model for longer scenes, multiple keyframes and mixed references. This guide covers the full model, the smaller Flash version, and how to prepare a shot. Kling 4.0 generation is not yet available on VideoDance.

Where you can use Kling 4.0 today

Last checked October 14, 2026. The official September 30 announcement describes early access and a wider October rollout. Flash began limited access for Ultra Yearly subscribers on September 28. We have not confirmed general access for every account.

VideoDance currently offers Kling 3.0. Use its page for a working generator and an actual credit quote. The 4.0 features described here belong to the announced model; those controls are not yet available on VideoDance.

Generate with Kling 3.0 on VideoDance

Kling 4.0, Flash and the version available here

Kling 4.0, Flash and the version available hereKling 4.0: announced full modelKling 4.0 FlashKling 3.0 on VideoDance
Clip length3–30 seconds3–20 seconds5, 10 or 15 seconds
Image controlFirst/last frames; up to 10 keyframesFirst frame onlyText-to-video on VideoDance
ResolutionOfficial specification lists up to 4K; channel limits apply720p720p or 1080p
Reference workflowOmni Reference and video editingOmni Reference; a subset of full capabilitiesThe site does not expose the full Omni workflow
AvailabilityEarly access / October rollout announcementLimited early accessAvailable on its model page

Read output specifications together with the access notes

The official 4.0 specification lists 4K, while the September 30 announcement marks 10-bit HDR at 1080p and 4K as coming soon. fal lists 720p and 1080p for its early-access channel. Confirm the export you need in the channel you will actually use.

Extension to two minutes is also announced as coming soon. It is a sequence of extensions, not a two-minute single generation. Flash is capped at 720p and 8-bit SDR, and cannot use a last frame or multiple keyframes.

What changes from Kling 3.0?

The official comparison adds longer generation, more reference inputs and intermediate keyframes. Audio and 4K already existed in parts of the 3.0 family. Calling either one a first for 4.0 would miss the practical upgrade: more ways to describe how a scene develops.

Choose by the job. A short portrait motion study needs fewer controls than a product reveal with a specified ending. If you require an intermediate visual state, check for keyframes. If a face or object must survive a camera change, check the reference mode. A model name alone does not tell you which controls a site exposes.

Choose the input for the shot

Start with text
Use text when the subject and setting are still open to interpretation. Describe one main action and its result. Decide whether you want a continuous take or several shots before adding camera directions.
Start with an image
Use an image when the opening composition matters: the face, package, lighting or product position is already approved. Describe what moves from that state. Repeating every visible detail can bury the action you want.
Start with references
Assign a purpose to each input. A character sheet establishes appearance; a location image establishes the space; a video can guide movement or camera work. Remove references that disagree about the same detail.
Start with existing footage
Editing is useful when a take has something worth keeping. Write down what should change and what should stay fixed, then inspect both. A successful replacement is incomplete if it also changes a product label you needed to preserve.

Build a reference set that agrees with itself

Kling lists up to 15 combined reference assets for the full model, with individual limits of 10 images, 5 videos and 7 subjects. Video references share a 30-second total. Those limits do not add up to 22 simultaneous inputs.

For a bottle demonstration, prepare a front view, a view of the cap and a clean image of the table. Give each file a role. If the cap is blue in one image and silver in another, resolve that conflict first. Keep the same label placement and proportions across the set.

A reference and a keyframe answer different questions. The reference says which bottle should appear. A keyframe says which visual state the scene should reach. Use the prompt for the action between states, such as a hand reaching for the cap.

Plan a 30-second product scene

  1. 1

    Set the final use

    Decide where the video will appear and what viewers should understand. For a simple product demonstration, choose one action that proves the feature. A vertical social clip and a wide presentation need different framing.

  2. 2

    Choose the important moments

    Sketch the opening, the interaction and the ending. Add an intermediate keyframe only when it communicates a necessary change. Ten available keyframes are a maximum, not a requirement.

  3. 3

    Match the references

    Check identity, clothing, product geometry and lighting before submission. When a change is intentional, specify it. Otherwise, keep those details consistent across your images.

  4. 4

    Describe the transitions

    Write what connects each moment: who moves, where the camera goes and what sound occurs. Avoid asking for a locked camera and an orbit during the same beat.

  5. 5

    Review the whole take

    Watch at normal speed, then inspect the transitions and contact points. Keep a record of the input, settings, attempts and actual charge. A good opening frame does not establish continuity through the ending.

An original storyboard to adapt

An original storyboard to adaptVisual milestoneDirection between milestones
0–6s: establishClosed bottle beside a backpack on a deskA gentle push-in; keep the bottle upright
6–12s: revealHand holds the cap above the open bottleOpen the cap once; keep contact believable
12–22s: demonstrateWater pours from the bottle into a cupKeep the cup on the desk and the stream connected
22–30s: finishClosed bottle returns beside the backpackSet it down, then hold a clean closing composition

Use the storyboard as a plan, not a timing guarantee

This timeline is our teaching example. It has not been generated with Kling 4.0. In a channel that exposes keyframe timing, set the milestones there and use the text to describe their connection. A timestamp in a prompt alone is not proof that an event will land at that exact second.

If opening the cap and pouring water create too much work for one take, split them into two scenes. Review the product in each scene before joining them. For a campaign that requires exact lettering, compare the label against the approved source and finish the text in an editor if needed.

Prompt exercise: one product, one action

Original teaching prompt; not a tested output. Use a matching first-frame image and confirm that the chosen channel supports the duration and sound settings.

Animate the supplied image as a 10-second tabletop product shot. Keep the lamp’s silhouette, hinge and matte finish consistent. A hand enters from frame right, tilts the lamp head once, then leaves. The base stays on the table. The camera makes one slow push-in, ending on the illuminated notebook. Quiet room ambience and one hinge click. No dialogue, music or added text.

Replace the lamp with your own object and name the part that moves. Inspect fingers, the hinge, the base and the shadow. When comparing variations, change the camera instruction or the action, rather than changing everything together.

Prompt exercise: connect the keyframes

Original teaching prompt for the full model’s keyframe workflow; Flash does not offer that control. Prepare the images before adapting the text.

Use the references to keep the same blue insulated bottle throughout. Follow the supplied keyframes: begin on the closed bottle beside a backpack, show a hand opening the cap, show water poured into a cup, and finish with the closed bottle in the original setting. Connect these moments with ordinary hand movements and deliberate camera cuts. Keep the bottle proportions, cap and lighting consistent. Natural handling sounds; no speech or new lettering.

Set the milestones in the interface when available. Review whether the hand, cap and water remain connected during transitions. If a reference shows an open cap while the matching keyframe shows it closed, resolve the input before trying again.

Prompt exercise: give each speaker a line

Original teaching prompt; not a verified lip-sync demonstration. Check the exact model’s audio controls first.

One medium shot in a quiet workshop. Alex places a repaired wooden toy on the table. Alex says in calm English, ‘It moves again.’ Sam looks at the toy, then says, ‘You kept every scratch.’ Only the assigned speaker talks during each line. Preserve both faces and clothing. Background workers continue their own tasks. Subtle room ambience, one wheel rattle, no subtitles or music. Keep the camera still.

Listen before judging the visuals. Check which character speaks, whether the wording survives and whether the pauses fit the performance. If the clip needs an exact script, allow for a separate audio edit.

Official creator examples: what to look for

Johnson Sheng’s October 2 article presents action, camera movement, prop continuity and dialogue cases. He is a Kling creative partner; these are vendor-published examples, not a neutral success-rate study.

Oasis’s October 10 article explores natural performance in a surreal bathroom scene and continuity across different images. The useful viewing question is whether character movement motivates the next scene, rather than simply displaying the images in sequence.

Open the original case articles in the sources below to view their examples. When examining any demo, separate a visible result from an unknown production process. Without the attempts and settings, one polished clip cannot tell you the likely cost or success rate for your own scene.

Early Flash feedback and a practical review checklist

In a Reddit post, Jack_P_1337 reported unwanted music despite a no-music instruction, cuts despite a continuous-shot request, and occasional element inconsistency. The same user praised motion and voice acting. These are early Flash observations from one person, not established defects in the full model.

For your own test, check one requirement at a time: an uninterrupted shot, a stable product, or two correctly assigned lines. Count failed attempts as well as successful ones. If you need a continuous take, watch for a cut that hides a change in the character. If silence matters, listen through the entire file.

Access, pricing and what VideoDance offers

fal’s page describes early access through its team. It is not proof of public self-service access for your account. Its demo clips are attributed to Kling AI, not to a fal API run. We have not confirmed a public 4.0 API price or completed a 4.0 request.

VideoDance has not set a Kling 4.0 credit price. Its existing Kling 3.0 page offers 5, 10 or 15 seconds at 720p or 1080p, with optional native audio and a credit quote before submission. Do not use that quote to estimate a 30-second 4.0 scene.

When pricing becomes available, compare the same duration, resolution and audio setting. Record what one usable clip actually costs across all attempts, rather than dividing a subscription fee by an assumed number of perfect results.

Kling 4.0 questions

Available models and reference-image creation

Sources and last checked

Maintained by VideoDance · Facts checked

This guide reviews public documentation. We have not generated or benchmarked Kling 4.0, or independently inspected the linked demo videos. The prompt exercises and timeline below are our own teaching examples, not tested outputs. Access and prices can change. The date below is in UTC (October 14 in Taipei).