[Bg2] Kling 3.0 is a refined AI video model supporting text-to-video and image-to-video. It generates native audio with natural lip alignment. The built-in director mode arranges multiple shots in one prompt, adjusting framing and camera movement. Reference anchoring keeps characters consistent across shots. It supports 4K output for creative clips and previsualization.