Coming soon
Kling 4.0: Multi-Keyframe Control and Omni Reference
Kuaishou’s new-generation Kling video model, coming soon to A2E. Pin the story down with up to 10 keyframes, hold characters steady with up to 15 references, and generate up to 30 seconds in one pass — with stereo sound made alongside the picture.
What Kling 4.0 puts out
Every clip below comes from Kling’s own launch demos, so you can judge the picture quality before spending anything. They loop silently here — the stereo sound is covered further down.
Six things it does well
The claims are Kling’s own and the clips are their demos. What you get on A2E is whatever your own generation returns.
Motion that holds, sound that matches
Dynamic Motions and complex camera moves stay stable: running, fighting and chase scenes, tracking and orbit shots keep the picture steady and continuous. Sound is generated with the picture — high-quality audio in two-channel stereo, with more accurate lip sync, so voices, dialogue and performance stay in step.
Omni Reference: up to 15 inputs in one job
Mix images, videos and elements, and give an element its own voice. Images anchor appearance, videos carry motion and camera movement, voices bind vocal identity, and elements run through the whole clip. One job takes up to 15 references — up to 10 images, up to 5 videos (30 seconds combined) and up to 7 elements — and they work together more reliably, keeping characters, actions, shots and audio closer to what you intended.
Multiple Keyframes: up to 10, shot by shot
Anchor several key visual moments in a single generation and specify character states, scene changes and story beats, so the clip unfolds the way you planned. Use 2 keyframes to fix the start and the end, 5 to control the turning points, or 10 to lay out a full storyboard sequence. Characters and scene style stay more naturally connected from one keyframe to the next.
Native 30s Generation: a whole beat in one pass
Generate up to 30 seconds in a single generation, with long takes and continuous storytelling that keep characters, scenes, voices and style consistent and cut down on stitching. Guide the pacing with time-based descriptions — or start from one sentence and let the model expand it into story, staging and camera work with a director’s eye.
Creative Recreation and Video Editing: build on what exists
Creative Recreation references the pacing, shot composition and expression of a video that already works, and swaps in your own products, characters and assets. Video Editing takes up to 5 clips, with 1 as the primary video, and makes targeted edits to characters, actions, backgrounds, style and camera while the rest of the scene stays stable — no regenerating the whole clip.
Languages, accents and on-screen text
Voices in Chinese, English, Japanese, Korean, Spanish, Portuguese, German, French and Hindi, with dialects such as Sichuan dialect and Cantonese and accents from Beijing and Northeastern Mandarin to American, British and Indian English. Text Generation covers the same nine languages plus emoji and logos, and the text stays clear and stable through camera moves and scene changes.
Kling 4.0 or Kling 4.0 Flash
Kling ships this generation in two versions. Which one you can pick on A2E is whatever the model selector in the generator shows.
Kling 4.0 — the full feature set
Text to Video, Image to Video, First & Last Frame, Multiple Keyframes, Omni Reference and Video Editing all live here. 3 to 30 seconds, 720p, 1080p or 4K, and aspect ratios that include 21:9 ultrawide. Use it for finished pieces and for shots you need to control precisely.
Kling 4.0 Flash — fast drafts
Built for frequent creation; Kling describes it as faster to generate and better value. It supports Text to Video, Image to Video from a first frame, and Omni Reference, at 3 to 20 seconds and 720p. Use it to find the direction, then switch to Kling 4.0 for the final.
Where it earns its place
The use cases Kling names for 4.0, and the capability each one leans on.
Film and short drama
AI films, short dramas, animated comics, storyboard previsualization and post-production touch-ups. Native 30s Generation carries a whole story beat, Multiple Keyframes steer where the plot goes, and Omni Reference keeps characters and voices consistent.
Advertising and e-commerce
Brand films, product showcases, performance ads and listings for cross-border stores. Creative Recreation carries a proven idea over to a new product, Text Generation puts the selling point and the brand message in the frame, and Video Editing produces the next version without starting over.
Short video and music video
Lifestyle clips, vlogs, talking-head pieces, music videos and lyric visuals. One image or one reference video is enough to start, and the same piece can be restyled for each platform.
Games and interactive entertainment
Character promos, cutscenes, update trailers and user-acquisition creatives. Character art, 3D renders, gameplay footage and voice assets all work as references, keeping a character’s look and voice more consistent across the clip.
What Kling 4.0 supports
These are the specifications Kling publishes for 4.0 — enough to judge whether it clears your delivery bar. The tiers A2E opens are the ones you can select in the generator.
| Setting | What you get |
|---|---|
| Duration | 3 to 30 seconds in one generation. |
| Resolution | 720p, 1080p or 4K. |
| Aspect ratio | 21:9, 16:9, 1:1, 9:16, or Auto. |
| Audio | Two-channel stereo, generated with the picture. |
| Keyframes | Up to 10 keyframe images. |
| References | Up to 10 images; up to 5 videos with a combined duration of 30 seconds; up to 7 elements, including up to 3 video-based elements; voice reference for audio. Up to 15 inputs in total. |
| Prompt | Up to 8,000 tokens — room for a shot-by-shot script, not just a sentence. |
| Video Editing | Up to 5 input clips, with 1 as the primary video. Each clip runs 3 to 30 seconds, 30 seconds combined. Kling recommends pairing it with reference images. |
| Coming soon | 10-bit HDR output at 1080p and 4K, and Video Extension — extending an uploaded video across multiple generations, up to 2 minutes in total. Kling lists both as coming soon. |
Four things that save a retry
Don’t start with all ten keyframes
Two fix the start and the end, five control the turning points, ten make a full storyboard. Settle the direction with a few, then add the ones in between — easier than hunting for the one frame that threw a full set off.
Give every reference a role
Images for appearance, videos for motion and camera movement, a voice for vocal identity, elements for whatever has to run through the whole clip. When a mixed-reference job goes wrong, it is usually two assets fighting over the same thing.
Write long clips against the clock
If several things happen inside 30 seconds, split them into timed segments in the prompt. An unsegmented description tends to be acted out in the opening seconds.
Short prompt or long prompt, on purpose
With only an idea, write one sentence and let the model expand the story and the camera work. With the shots already decided, spell them out and it follows. Short for surprise, long for control.
Docs and API
Kling 4.0 joins the A2E developer docs when it goes live. Endpoints, parameters, authentication and billing for every model are documented there, and MCP access is available too.
Why run Kling 4.0 on A2E
Runs in the browser
No GPU, no install and no API key. Kling 4.0 draws on your A2E account like every other model on the site.
Every model, one account
Kling, Wan, Seedance, MiniMax, Sora and Veo sit behind one account and one balance. Run the same brief through several and keep the best.
From clip to finished asset
Lip sync, voice cloning, face swap, subtitle removal and upscaling live next to generation, so a clip can go from prompt to finished asset without leaving the site.
FAQ
What is Kling 4.0?
Kling 4.0 is Kuaishou’s new-generation Kling video model. It generates video from text, images, first and last frames or up to 10 keyframes, accepts up to 15 references in one job, runs up to 30 seconds in a single generation and produces stereo sound with the picture.
Kling 4.0 or Kling 4.0 Flash — which should I use?
Flash is built for frequent creation: faster and better value, at 3 to 20 seconds and 720p, without first-and-last-frame or keyframe control. Kling 4.0 is the full model, up to 30 seconds and 4K. Draft in Flash, finish in Kling 4.0. The versions you can pick are shown in the generator.
How many references can I use at once?
Up to 15 in one job: up to 10 images, up to 5 videos with 30 seconds combined, and up to 7 elements, of which up to 3 can be video-based. An element can carry its own voice reference.
Does Kling 4.0 generate sound?
Yes. Kling 4.0 outputs two-channel stereo generated with the picture, with voices in nine languages plus a range of dialects and accents. Describe dialogue and sound in the prompt alongside the action.
Are 10-bit HDR and Video Extension available yet?
Kling lists both as coming soon: 10-bit HDR output at 1080p and 4K, and Video Extension up to 2 minutes. They will appear in the generator once Kling opens them.
Is Kling 4.0 free to use?
You can start on A2E for free: new accounts get bonus credits on signup, no credit card is required, and you can claim extra credits every day by checking in. What a Kling 4.0 generation costs is shown in the generator before you submit.
Can I use the videos commercially?
Yes. Videos generated on a paid A2E plan can be used commercially — ads, social media, client deliverables and e-commerce. You keep ownership of the output, with no watermark, no attribution requirement and no per-video royalties.
Where do my videos go after generation?
Every finished job appears in the My Results tab beside the generator, where you can preview, download, share or delete it. The list keeps the last 3 days of generations, so download anything you want to keep.
Related AI video models on A2E
- Wan 3.0 — Tongyi Wanxiang’s all-in-one video model, with a native 30-second single take and documents turned straight into video.
- Seedance 2.5 — ByteDance’s professional multimodal model, also live on A2E. Compare the two on the same brief.
- MiniMax H3 — omni-modal generation with native stereo sound, at up to 2K.
- Kling 3.0 and Kling 2.6 — earlier Kling models, still available on A2E.
- Lip sync and video to audio — finishing tools for generated footage.