Kling 4 AI video generator for text-to-video and image-to-video with native audio, running the newest Kling engine open to third-party sites. Kling 4.0 goes live here the day its API opens.
Kuaishou Kling AI model overview
Kling 4, officially written Kling 4.0, is the latest generation of Kuaishou's Kling AI video model family. It follows Kling 1.0 (June 2024), the 2.x series and Kling 3.0, and Kuaishou announced it on September 28, 2026. The Kling 4 AI model generates video from text, images and reference clips, and the lineup also includes a lighter Flash version built for fast, frequent work.
The official specifications: 3-30 second clips at 720p, 1080p or 4K, 10-bit HDR on the 1080p and 4K tiers, up to 10 keyframes, two-channel stereo audio with lip sync across nine languages and a range of accents, prompts up to 8,000 tokens, and omni reference that accepts up to 15 items in total, including up to 10 images and 5 reference videos. Flash renders 3-20 second clips at 720p. Video extension to two minutes is listed as coming soon. The full launch is set for October; Flash opened to limited early access on September 28.
Capabilities
Describe a shot in plain English: subject, action, camera move, lighting and sound. The generator turns it into a cinematic clip in minutes and follows long, multi-part prompts closely. The next model is reported to add longer shots and finer camera control.
Upload a product photo, portrait or storyboard frame and use it as the first frame. The model animates the scene while keeping faces, logos and composition locked to your image, which is ideal when the look is already approved and you only need motion.
The Kling 4 AI model accepts up to 10 keyframe images, so you set the exact beats a shot has to hit instead of rerolling a whole clip. First-and-last-frame control is supported too, and clips run up to 30 seconds natively, long enough for a complete action rather than a fragment.
Omni reference takes up to 15 items in one prompt: up to 10 images, up to 5 reference videos totalling 30 seconds, voice references and up to 7 subjects. Use them to lock a face, an outfit, a camera language or a performance rhythm. Upload only media you are allowed to use.
Two-channel stereo audio is generated with the picture, with lip sync across Chinese, English, Japanese, Korean, Spanish, Portuguese, German, French and Indonesian, plus regional accents. It also renders readable on-screen text and emoji in those languages, which earlier generations struggled to keep stable.
4.0 vs 3.0
A planning comparison, not a benchmark. Every figure comes from Kling's own release notes and API docs, checked on September 29, 2026. The Kling 4 AI model was announced on September 28 and launches in full in October.
| Feature | Kling 4.0 | Kling 4.0 Flash | Kling 3.0 | Kling 3.0 Turbo |
|---|---|---|---|---|
| Release status | Announced, Oct 2026 | Early access, Sep 28 | Live, Feb 2026 | Live, Jun 2026 |
| Clip length | 3-30s | 3-20s | 3-15s | 3-15s |
| Max resolution | 4K, 10-bit HDR | 720p, 8-bit SDR | Native 4K | 1080p |
| Keyframe control | Up to 10 keyframes | — | Start & end frame | — |
| Omni reference | Up to 15 items | Yes | Images or a video | — |
| Native audio | Stereo, 9 languages | Yes | Yes, 5 languages | Always on |
| Prompt length | 8,000 tokens | 8,000 tokens | 2,500 chars | 2,500 chars |
| Video extension | Up to 2 min, coming soon | — | — | — |
| Aspect ratios | 21:9 to 9:16 | 21:9 to 9:16 | 16:9, 9:16, 1:1 | 16:9, 9:16, 1:1 |
| Available here | At API launch | At API launch | Yes | Yes |
Bottom line: Kling 3.0 is what you can generate through the API today, and it already gives you native 4K and native audio. The Kling 4 AI upgrade is about length and control: 30-second clips, 10 keyframes, 15 reference items and 10-bit HDR. Prompts you refine now carry straight over to it.
Workflow
No app to install and no waitlist for the current engine. Open kling4.io in any desktop or mobile browser and go from idea to download in three steps.
Choose Text to Video or Image to Video. Lead with the subject, then add the action, camera move, lighting and sound. For image-to-video, upload a sharp JPG, PNG or WebP.
Pick the duration, resolution and aspect ratio: 16:9 for YouTube, 9:16 for TikTok and Reels, 1:1 for feeds. Switch native audio on or off. The exact credit cost appears before you generate.
Most clips render in one to three minutes. Preview the result, download the MP4 and reuse the prompt to iterate. If a render fails, your credits are refunded automatically.
Use cases
A 15-second product ad used to mean a studio, a crew and a week of edits. Many AI tools cut the cost but break the shot: faces drift, hands melt, audio is missing. Kling models are built for exactly those weak spots, so one prompt here can become a clip you can publish.
Turn a single packshot into a rotating hero shot, an unboxing or a lifestyle scene. Test three ad angles for Amazon, Shopify or TikTok Shop before paying for a photo shoot, then keep the winner.
Block scenes, test camera moves and pitch a mood before the shoot. Multi-shot generation keeps characters consistent between cuts, so a storyboard reads like a sequence rather than a mood board.
Batch vertical 9:16 hooks for TikTok, Reels and YouTube Shorts, with voice and sound in the same render. Generate several variations, publish the strongest and iterate on the prompt the same day.
Animate history scenes, science concepts and product tutorials that would be slow or expensive to film, and add narration in your audience's language with native audio.
Start with free credits, then choose a monthly plan or a one-time credit pack. Every option shows how many seconds of video it covers before checkout, failed renders are refunded, and credits you buy today will work with the new model at launch.
Choose Subscription Plan
Perfect for individuals and light users
Billed monthly · cancel anytime
INCLUDES
For professional creators and teams
Billed monthly · cancel anytime
INCLUDES
Designed for large enterprises and professional studios
Billed monthly · cancel anytime
INCLUDES
To cancel your subscription, please contact support: support@kling4.io
FAQ
Straight answers about the release date, official specs, pricing and this independent site.
Write one clear scene, choose text-to-video or image-to-video, and judge picture and sound together. You start with free signup credits, they carry over to the Kling 4 AI model at launch, and there is nothing to install.
