Sample
Kuaishou
video model

Kling v3

A unified multimodal video model tuned for physically accurate motion — human movement, gestures, and dynamic action avoid the stiff, uncanny feel that trips up lesser video models, with cinematic camera control built in.

280 cr/secText promptReference imageReference video

Excels at

  • Strong human-motion realism (dance, sports, dynamic action)
  • Cinematic camera movement control
  • Native audio generation alongside video

Signal

Use when

  • Fashion and product motion clips
  • Dance, sports, and dynamic action sequences
  • Shots needing precise, described camera movement

Avoid when

  • Long multi-shot narrative sequences (see Wan 2.7)
  • Budget-constrained high-volume video generation

Prompt used for this sample

A paper boat drifting down a rain-slicked city street at dusk, gentle camera follow, cinematic