Chinese artificial intelligence startup MiniMax has introduced the video-1 algorithm, which generates short video clips based on text prompts. The video-1 generator was presented to the general public at the company’s first developer conference in Shanghai a few days ago, and later became available to everyone on the MiniMax website.

Image source: scmp.com

With video-1, the user can create videos of up to 6 seconds based on a text description. The process of creating such a video takes about 2 minutes. MiniMax founder Yan Junjie said at the presentation that video-1 is the first version of the video generation algorithm based on text prompts, noting that in the future the neural network will be able to create videos based on static images, and will also allow editing already created clips.

The emergence of video-1 reflects the desire of Chinese technology companies to make inroads into the nascent segment of the AI ​​market. The video generator was introduced just a few months after the announcement of OpenAI’s Sora neural network, which also allows you to create videos based on text prompts. As for MiniMax, the company was founded in December 2021 and has done quite a bit of work since then. The new video-1 tool is offered as part of the MiniMax platform called Hailuo AI, which is aimed at the consumer market and already provides access to text and music generation functions using neural networks.

In addition to MiniMax, other Chinese companies are also developing AI algorithms for generating video from text. Beijing startup Shengshu AI launched its own video generator from text in Chinese or English called Vidu in July. The $1 billion-plus startup Zhipu AI that same month unveiled its Sora, which can create short videos based on text prompts or static images.

The owner of TikTok and Douyin, ByteDance, last month published the Jimeng text-to-video application for generating videos from text in the Chinese App Store, and even earlier it appeared in local Android application stores. Jimeng allows you to create 80 images or 26 videos for free, and for more active interaction with the neural network, a subscription is offered for 69 yuan (about $10). Last month, Alibaba Group Holding announced the development of a video generation algorithm called Tora, based on the OpenSora model.

Note that MiniMax’s investors include large IT companies such as Alibaba, Tencent Holdings and miHoYo (creator of Genshin Impact). The next round of financing took place in the spring and after its completion, the market value of MiniMax was estimated at more than $2 billion.

Leave a Reply

Your email address will not be published. Required fields are marked *