Kuaishou's next Kling video model is built for visual realism, creative control and complete stories.
Dark fantasy cinematic CG scene in a ruined temple beneath storm clouds. The silver-haired female warrior in 
@Image1 wears torn white battlewear and is bound by glowing purple magic chains. 
@Image2 A distant throne mage casts spells before a giant violet rune circle as broken stones float in the air. The camera cuts from a villain close-up to the warrior falling, struggling, lifting her head, awakening, and defeating the enemy with magic. Oppressive mood, epic scale, dramatic lighting. Use music.mp3 as the music reference.
59 sample videos from Kling AI's Kling 4.0 release note, grouped the way Kling presents them. Hover or tap a card to play it.
Stable large-scale motion, immersive sound and output closer to a finished professional cut.
Generated from a reference image, in the new 21:9 ultra-wide format.
Generated from keyframes. Large, fast movement stays stable and continuous.
Generated from a first-frame image, with fast action and a big camera move.
Generated from a first-frame image. Dialogue, sound and performance stay in sync.
Generated toward a last-frame image, with native audio.
Generated from keyframes, with dual-channel stereo audio.
Generated from a first-frame image. Small glances and hesitations read clearly on camera.
Generated from a text prompt.
PromptA vlogger video: the woman happily introduces herself to her audience. Keep it feeling real, with realistic image quality.
Rendered in 10-bit HDR. Shown here converted to standard dynamic range for the web.
Rendered in 10-bit HDR. Shown here converted to standard dynamic range for the web.
All example videos were made by Kling AI with Kling 4.0 and published in the official Kling 4.0 release note on September 28, 2026. Captions are translated from the Chinese release note, and the videos are re-encoded for the web. View the official release note
Kling built Kling 4.0 around four upgrades: the picture and sound, how you reference and edit, how long a story can run, and how many languages it can speak. Each card plays an official Kling 4.0 example.
Kling 4.0 generates up to 30 seconds natively, so a scene with a setup, a turn and a payoff can come out of one generation instead of stitched clips.
Multi-keyframe control lets you pin down the moments a video has to hit, with up to 10 keyframes guiding one continuous shot sequence.
Omni Reference takes up to 15 references per task: up to 10 images, 5 videos (30 seconds in total) and 7 subjects, with voice references for subjects. The same inputs power video editing.
Output goes up to 4K, with 10-bit HDR on 1080p and 4K (coming soon), a new 21:9 ultra-wide format, dual-channel stereo audio and more accurate lip sync.
Kling 4.0 is the next generation of Kling AI, the video model family developed by Kuaishou. Kling introduced it on September 28, 2026 with the "Meet KLING 4.0 | You Call the Shots" trailer, describing a model that takes video creation to a new level of visual realism, creative control and narrative completeness.
Stable dynamic motion, high-quality stereo audio, more accurate lip sync, up to 4K resolution and 10-bit HDR output.
Richer reference options with up to 15 multimodal references, more consistent results and enhanced video editing.
Multi-keyframe control with up to 10 keyframes and native 30-second generation.
Video extension, plus support for multiple languages, accents and dialects.
Combine a prompt with up to 15 multimodal references and up to 10 keyframes to direct characters, look and story beats across a native 30-second video.
Specs as announced by Kling. Exact controls on Topview will be confirmed at launch.
Kling 4.0 adds enhanced video editing through Omni Reference and supports video extension, so an existing shot can be changed or continued instead of regenerated from scratch.
Specs as announced by Kling. Exact controls on Topview will be confirmed at launch.
Collect the faces, products, locations, motion clips or voices you want to keep consistent. Kling 4.0 supports up to 15 multimodal references.
Write the prompt and place up to 10 keyframes for the moments that matter. Kling 4.0 can generate up to 30 seconds natively.
Use video editing to change a specific part of a shot, or extend the video to keep the story going, then export up to 4K.
Longer native clips, keyframes and heavy referencing matter most to people who tell stories in sequences, not single shots.
Plan beats with up to 10 keyframes and let a single 30-second generation carry a whole scene.
See moreKeep a product, model and setting consistent with multimodal references across every version of a spot.
See moreStereo audio and more accurate lip sync make performance shots and music-driven edits easier to plan.
See moreDirect a campaign film with references and keyframes, then edit or extend shots instead of starting over.
See moreNative 30-second scenes and consistent characters help episodic stories hold together from shot to shot.
See moreKling 4.0 supports multiple languages, accents and dialects for versions made for different markets.
See morePrompts Kling published with its Kling 4.0 examples, next to the videos they produced.
A vlogger video: the woman happily introduces herself to her audience. Keep it feeling real, with realistic image quality.
Translated from the Chinese release note.
An old man ranting about AI.
Original English prompt.
A 7-second continuous shot. In the UK, two friends who haven't seen each other in a long time meet at the corner of a quiet street at dusk, with all kinds of people going about their lives around them. They smile, walk quickly toward each other and hug warmly. Natural expressions and body movement. Handheld camera, slightly shaky, following them. Soft warm light, nostalgic film look. Natural footsteps, street ambience and background music; they speak English.
Translated from the Chinese release note.
How the announced Kling 4.0 specs compare with Kling 3.0, which Kuaishou launched in February 2026.
Kling 3.0 details from Kuaishou's February 5, 2026 launch announcement. Kling 4.0 details from Kling AI's September 28, 2026 announcement.
From Kling AI's Kling 4.0 trailer and release note. Some features are marked by Kling as coming soon.
Next-generation Kling video model developed by Kuaishou.
Kling 4.0 Flash opened for a limited group on September 28 and is live for Kling Ultra Yearly subscribers.
Kling says Kling 4.0 is coming this October.
Native single-pass generation up to 30 seconds. Video extension up to 2 minutes is coming soon.
10-bit HDR on 1080p and 4K is coming soon. Aspect ratios: 21:9, 16:9, 1:1, 9:16 or smart.
Up to 10 images, 5 videos (30s total) and 7 subjects per task.
Plus text-to-video, image-to-video and first/last frame.
Dual-channel stereo audio with more accurate lip sync.
Chinese, English, Japanese, Korean, Spanish, Portuguese, German, French, Hindi and more, plus accents such as American, British and Indian English.
Prompts up to 8,000 tokens. Topview pricing and availability will be confirmed at launch.
Native 30-second stories, 10 keyframes, 15 references and 4K HDR. We'll switch this page on as soon as Kling 4.0 generation is available on Topview.
Specs on this page come from Kling AI's official announcement and may change before launch.