Kuaishou launched Kling 3.0 on February 5th, and the jump from earlier versions is striking. Where Kling 2.6 was limited to single continuous shots, the new model introduces multi-shot storyboards — up to six camera cuts in a single generation. Video duration extends to 15 seconds with custom timing.

The headline features that matter for creative work: an Elements system that maintains character identity across shots, three-speaker dialogue with individual voice tracking, and support for five languages including English, Japanese and Korean. The multi-shot storyboard lets you specify duration, shot size, perspective and camera movements for each segment, which turns what was essentially a clip generator into something closer to a production tool.

Against the current competition — Runway Gen-4.5, Veo 3.1, Seedance 1.5 Pro — Kling 3.0 leads on resolution and multi-shot capability, though Runway still edges ahead on overall quality for certain styles and Seedance has the tightest lip-sync precision for dialogue.

The pace of advancement in this space over the past eighteen months has been remarkable. What took hours of manual compositing in 2024 now generates in seconds. The gap between AI-generated video and professional footage continues to narrow with each model release.

I scanned an image of fashion model Gail Elliott from a 1992 Spring/Summer Escada catalogue and fed it to Kling 03 Pro with a custom prompt. It generated 15 seconds of video with audio from a single still.

A 1992 Scan Learns to Move