Kling

Large-scale reasonable movement Kling uses a 3D spatiotemporal joint attention mechanism to better model complex spatiotemporal movement, generate video content with large-scale movement, and conform to the laws of movement. Video generation up to 2 minutes Thanks to efficient training infrastructure, extreme reasoning optimization and scalable infrastructure, Kling's large model can generate videos up to 2 minutes long with a frame rate of 30fps. Simulate physical world characteristics Based on the powerful modeling capabilities inspired by the self-developed model architecture and Scaling Law, Kling can simulate the physical characteristics of the real world and generate videos that conform to the laws of physics. Powerful concept combination capabilities Based on a deep understanding of text-video semantics and the powerful capabilities of the Diffusion Transformer architecture, Kling can transform users' rich imagination into concrete pictures and fictional scenes that will not appear in the real world. Movie-level image generation Based on the self-developed 3D VAE, Keling can generate movie-level videos with 1080p resolution, which can vividly present both the vast and magnificent grand scenes and the delicate close-up shots. Supports free output video aspect ratio Keling adopts a variable resolution training strategy, which can output a variety of video aspect ratios for the same content during the inference process, meeting the needs of using video materials in richer scenes. Expression and body drive Based on the self-developed 3D face and body reconstruction technology, combined with background stability and redirection modules, the expression and body full drive technology is realized. With only a full-body photo, you can experience the vivid "singing and dancing" gameplay.