You may have seen the news about OpenAI launching a text-to-video AI model Sora earlier this year. Sora is not even available to the public as of yet, and it’s already facing stiff competition.
The video-generation model that is making headlines for its impeccable outputs is made by Kuaishou Technology, a Chinese company that is known for its short video platform. The text-to-video generation model by Kuaishou Technology is called Kling, and it was developed by the LLM team at Kuaishou.
Kling reportedly employs technology similar to Sora to create 1080p HD videos with a maximum run time of two minutes while Sora has only been able to generate videos that are a minute long. It can create large-scale realistic motions that essentially mimic features of the real world. Kuaishou Technology has shared plenty of demonstration videos on Chinese social media networks. The visuals demonstrate Kling’s mastery in the field of artificial intelligence by creating an in-motion replica of a text instruction. A few of these videos have also appeared on X.
One of Kling’s text-to-video depictions shows a Chinese man sitting at a table and eating noodles with chopsticks. Another visual shows moving landscapes as seen through a train window. The videos are nothing short of hyper-realistic and It’s sending the internet into a frenzy.
Users can create videos in different aspect ratios with the Kling model, which uses advanced 3D face and body reconstruction made possible by the company’s proprietary technology, 3D VAE. Variable resolution training makes this feasible by enabling complete expression and limb motion from a single full-body image.
It should be mentioned that more Chinese text-to-video models exist. A text-to-video model called Vidu AI that can produce 16-second, 1080p videos was also released in China in around April. Nevertheless, Kling appears to be a testament to China’s rapid development of AI models and has definitely managed to put China on the map as a force to be reckoned with when it comes to AI advancements.
- Apple has Unveiled an Open-Source LLM Model - July 31, 2024
- Anthropic Has Released Claude 3.5 Sonnet to Rival GPT-4o and More - July 1, 2024
- China’s Text-to-Video AI Tool Emerges as a Competitor to Sora - June 24, 2024