Zhipu Launches AI-Powered Video Generator in Bid to Rival OpenAI’s Sora

00:00

00:00/00:00

Listen to this article 1x

A video created with Sora is displayed on a smartphone with the OpenAI logo visible in the background in Brussels, Belgium, on Feb. 16. Photo: VCG

Beijing-based large language model startup Zhipu AI launched its artificial intelligence-powered video generator Qingying on Friday, joining an already crowded and competitive sector in China, where developers are racing to create applications to challenge OpenAI’s Sora.

Using technology similar to Sora, Qingying is capable of generating videos with a resolution of 1440x960 pixels that are up to six seconds long with a text or graphic prompt, said Zhipu in a statement on its official WeChat account.

You've accessed an article available only to subscribers

Subscribe today for just $.99.

VIEW OPTIONS

Download our app to receive breaking news alerts and read the news on the go.

Get our weekly free Must-Read newsletter.

Share this article

Open WeChat and scan the QR code

DIGEST HUB

Digest Hub

Back

Explore the story in 30 seconds

Zhipu AI launched its AI-powered video generator Qingying, which offers 1440x960 resolution videos up to six seconds long based on text or graphic prompts.
Qingying uses technology similar to OpenAI’s Sora, aiming to reduce video generation costs and time, but remains far from commercial application.
Other Chinese AI models like Dreamina, Vidu, and Kuaishou's text-to-video models offer various video duration and quality features, highlighting the competitive nature of the sector.

AI generated, for reference only

Who’s Who

Zhipu AI: Zhipu AI, a Beijing-based startup, launched Qingying, an AI-powered video generator, capable of creating 1440x960 resolution videos up to six seconds long from text or graphic prompts. Built on Zhipu’s CogVideoX model, which utilizes Sora’s algorithm and DiT architecture, Qingying is trained with public and partner data and can generate dynamic emojis, adverts, and story plots. The company aims to reduce video generation costs and time for commercial viability.

Bilibili: Bilibili is a partner providing data to train Zhipu AI’s video generator, Qingying. It is a streaming platform, likely contributing to the diverse and extensive data sets used to enhance the AI's video generation capabilities.

ByteDance Ltd.: ByteDance Ltd., the parent company of TikTok, owns a web-based AI platform called Dreamina. In March, Dreamina unveiled a text-to-video model capable of generating footage from three to six seconds long, highlighting the company's involvement in the race to develop AI-powered video generation technologies in China.

Shengshu Technology: In April, Shengshu Technology, a Beijing-based generative AI startup, and Tsinghua University unveiled Vidu, an AI model capable of generating high-definition videos up to 16 seconds long.

Kuaishou Technology: Kuaishou Technology, a Beijing-based short-video specialist, launched its text-to-video model in June, capable of generating high-definition videos up to two minutes long. The model has undergone two major upgrades to enhance motion accuracy, composition, and color tone. Kuaishou began charging users to subscribe to the model earlier this week.

AI generated, for reference only

What Happened When

February 2024：: ChatGPT developer OpenAI unveiled Sora, sparking a rush among Chinese tech companies to develop their own versions.

March 2024：: Dreamina, a web-based AI platform owned by ByteDance Ltd., unveiled a text-to-video model that can generate footage from three to six seconds long.

April 2024：: Tsinghua University and Beijing-based generative AI startup Shengshu Technology unveiled Vidu, which can generate high-definition videos lasting up to 16 seconds.

June 2024：: Beijing-based short-video specialist Kuaishou Technology launched its text-to-video model that can generate high-definition videos that are up to two minutes long.

Before July 26, 2024：: Kuaishou Technology started charging users to subscribe to its text-to-video model, which has undergone two major upgrades to improve motion accuracy, composition, and color tone since its launch.