Kling 2.6

Kling Video 2.6 was released by Kuaishou's Kling AI on December 3, 2025. Its milestone capability is simultaneous audio-visual generation: from a text prompt or an image plus prompt, the model produces a short video with visuals, natural voiceover, sound effects, and ambient atmosphere in a single inference pass, replacing the traditional silent-footage-then-dub workflow.

The model supports text-to-audio-visual and image-to-audio-visual generation with native Chinese and English voice generation, multi-character dialogue, singing and rap with lip sync, physics-matched sound effects, and mixed soundtracks. Clips run up to 10 seconds at 1080p. It is a closed, API-served model aimed at creators, marketers, and educators producing short-form video quickly.