Picsart adds Kling 4.0 with 30-second, 4K video generation

Two models, two ways to start

Kling 4.0 is the newest video model in the Kling family. It generates video from a text prompt or an image and, according to the source, produces anywhere from 3 to 30 seconds natively in one pass, set in 1-second steps. Multi-shot is on by default, so the model splits a long generation into shots and paces them to fit the content.

Kling 4.0 Omni is the reference-driven companion. It is aimed at consistent characters, product swaps and edits to existing footage. The wider lineup also includes Kling 4.0 Flash, a lighter version for quick drafts and high-volume work. It produces clips of 3 to 20 seconds at 720p, trading length and resolution for faster turnaround at a lower cost.

• Kling 4.0: text prompt or image, 3 to 30 seconds, up to 4K, up to 10 keyframes.
• Kling 4.0 Omni: up to 15 references, 3 to 30 seconds, up to 4K, up to 7 consistent subjects, editing across up to 5 clips in one pass.
• Kling 4.0 Flash: 3 to 20 seconds at 720p, no keyframes.

Keyframes, camera work and prompts

To keep long clips on track, Kling 4.0 supports Multiple Keyframes. Users can place up to 10 keyframe images along the timeline to fix character poses, scene changes, product reveals and shot transitions, then let the prompt describe the action and camera movement in between. First and last frame control remains for simpler moves.

The model accepts prompts of up to 8,000 tokens, enough to script action, dialogue, camera moves and sound in detail. A one-line idea also works, and Kling 4.0 expands it into a fuller scene with story beats, blocking and camera movement. Aspect ratios include 21:9, 16:9, 1:1 and 9:16, or Auto, and output is available in 720p, 1080p or 4K.

The source says motion improvements show up in fast fights, sports, large-scale action and complex camera work such as tracking shots and orbits, where limbs, spatial relationships and camera rhythm stay connected.

References, editing and audio

A single Kling 4.0 Omni generation accepts up to 15 reference assets: up to 10 images, up to 5 videos with a combined length of 30 seconds, up to 7 subjects (up to 3 built from video) and up to 3 voice profiles. The idea is to give each reference a clear job, such as one image for face and outfit, another for the room, and a video for choreography or camera path.

Omni can also recreate a video by borrowing its pacing, shot composition and camera movement while swapping in new characters, products and settings. Targeted editing changes one element and leaves the rest of the scene alone, for example swapping a background, restyling a clip or removing a person from the foreground. An editing task takes up to 5 clips, including 1 primary video, each 3 to 30 seconds and 30 seconds combined. Pairing an edit with reference images gives the best results, according to the source.

Kling 4.0 generates two-channel stereo audio with sharper lip sync. Speech covers Chinese, English, Japanese, Korean, Spanish, Portuguese, German, French, Hindi and more, with accents and dialects such as Cantonese, Sichuanese, American, British and Indian English. Multilingual text, logos and emojis are meant to stay legible as the camera moves.

What it means for Picsart users

Kling 4.0 is available now in Picsart AI Playground, alongside Kling 4.0 Omni. Users pick a model, write a prompt or add references, and generate. Picsart points to brand ads, short films, social content, animation and music videos, and pre-production concept videos as use cases. The source says the release builds on Kling 3.0, with 30-second native generation, keyframes, 21:9 framing, stereo audio and German, French and Hindi speech listed as new in this generation.