Ghostboard pixel

Subscribe to Our Newsletter

Success! Now Check Your Email

To complete Subscribe, click the confirmation link in your inbox. If it doesn't arrive within 3 minutes, check your spam folder.

Ok, Thanks

Sora competitor Kling caused the internet to go into a frenzy

Kling, a new text-to-video generation model developed by Kuaishou Technology, gained significant attention for its impressive capabilities, sparking demand for state-of-the-art video generation tools and potential competition among leading AI companies.

Ellie Ramirez-Camara profile image
by Ellie Ramirez-Camara
Sora competitor Kling caused the internet to go into a frenzy
A small white rabbit wearing glasses sits on a chair in a café reading a newspaper, with a cup of hot coffee. Credit: Kling

Due to its impressive video generation capabilities, a new text-to-video generation model has gained traction for the last couple of weeks. Kling was developed by Kuaishou Technology, the maker of Kuaishou/Kwai, a popular TikTok competitor in China that boasts 400 million daily active users (DAUs). Kling appears to be in a public beta for Chinese users through the Kuaishou, Kwai, and KwaiCut apps, as many of those looking to access Kling outside China reported the apps required a Chinese phone number before granting users access to the model.

Regardless, the model keeps drawing attention on social media. Kuaishou Technology seems to be readying itself to share Kling with the world, as evidenced by the fact that it recently started offering an English translation of the model's official website. (The instructions to access the model remain the same and are still unavailable in English at the time of writing.) Meanwhile, the video generations using Kling, including a healthy amount of cat videos, keep steadily surfacing on X under the #Kling hashtag.

The official website states that Kling uses 3D spatiotemporal attention modules to improve the consistency of the relations between the elements in a video. The model outputs accurate, lifelike videos that conform to the laws of physics which are up to two minutes long at a 30 fps rate. The model also displays a strong semantic understanding, combining concepts drawn from short and straightforward, to more detailed and complex text prompts. Kling can also output videos with a 1080p resolution and several aspect ratios. More controversially, Kling also incorporates 3D facial and body reconstruction technology, enabling it to deliver convincing animations of full-body pictures.

Ever since OpenAI teased Sora, its closed-beta video generation model, professionals and enthusiasts alike have been waiting to try out the next big thing generative AI can offer. Kling's arrival on social media has demonstrated just how significant the demand for state-of-the-art video generation tools is. Undoubtedly, Kling's presence in the market will pressure OpenAI, Pika, and Runway to step up their offerings.

Ellie Ramirez-Camara profile image
by Ellie Ramirez-Camara
Updated

Data Phoenix Digest

Subscribe to the weekly digest with a summary of the top research papers, articles, news, and our community events, to keep track of trends and grow in the Data & AI world!

Success! Now Check Your Email

To complete Subscribe, click the confirmation link in your inbox. If it doesn’t arrive within 3 minutes, check your spam folder.

Ok, Thanks

Read More