Draft:Kling AI

Kling AI is a diffusion transformer text-to-video model that was created by the Chinese company Kuaishou Technology and later announced on June 6, 2024. Kuaishou claimed that the model can generate up to two minutes of video at 30 frames per second and in 1080p resolution. The large language model (LLM) uses three-dimensional face and body reconstruction using the company's proprietary 3D VAE (so-called "3D spatiotemporal joint attention mechanism" ) and the user can create videos in various aspect ratios.

Following KwaiYii and Kolors, the model's supposed video output was showcased on their website, and, the model is in beta access. The website also showcases a supposed demo of the "3D face and body reconstruction" technology's manipulation of a whole-body photo.

The model has been compared to that of OpenAI Sora text-to-video model; Kling AI has been deemed to be the "rival to Sora".