This Chinese AI video generator can take on OpenAI's Sora
What's the story
Chinese tech firm Kuaishou Technology has unveiled the upgraded version of its Kling AI model.
The company claims the new model is "the world's most powerful" in the realm of video generation.
It has already drawn over 22 million users from around the world, who have created over 168 million video clips and 344 million images, collectively.
The Kling AI 2.0 competes with OpenAI's Sora and Google DeepMind's Veo 2.
Background
A history of Kling AI
Kuaishou's new model comes just months after it introduced its previous generation of Kling models at the end of last year.
As per Artificial Analysis, a third-party service that tests and ranks global AI models, Kuaishou's previous-generation models are ranked as the world's top image-to-video and the second-best text-to-video models, behind Google's Veo 2.
Kuaishou is one of the many Chinese tech companies, including ByteDance and Alibaba Group Holding, racing to develop sophisticated AI tools to create high-quality videos.
Upgrades
Improvements in Kling AI 2.0
Kuaishou's Senior Vice President Gai Kun highlighted several improvements in the upgraded model, including improved instruction-following and prompt understanding capabilities.
The quality of images and movements has also been improved, along with a more realistic and aesthetic feel for the generated clips.
The videos generated by Kling AI 2.0 now feel more "punchy."