Recently, domestic AI video generation products such as MiniMax Conch AI and Kuaishou Keling AI have rapidly become popular overseas. On overseas social media such as
Most bloggers have nothing but praise for China's AI video generation applications, saying that the videos they generate are of higher quality, more stable and smoother. Some users even said that if it were not for some small details and bugs, they would even believe that these videos were shot in real life.
Recently, the iteration of domestic AI video generation applications continues to be hot. Zheng Lei, an analyst at Pacific Securities, believes that since the second half of the year, domestic AI video models have gained significant advantages.
In early September, MiniMax released the video model and launched it on Conch AI. When it was first released, the video model ranked first overall in the VBench test, a video generation model evaluation system jointly proposed by Shanghai Artificial Intelligence Laboratory and others. It currently ranks second, and is still in a leading position in terms of image quality and coherence.
Since the launch of the video model, Conch AI web version has received 4.97 million visits in September, a year-on-year increase of 867.41%, ranking first in the global growth rate list and the domestic growth rate list.
On September 19, Kuaishou Keling AI ushered in a major upgrade. The new Keling 1.5 model for video generation has significantly improved image quality, dynamic quality, aesthetic performance, motion rationality, and semantic understanding. On September 30, Kuaishou Keling’s AI video model product released a product update, focusing on the new “lip sync” function.
Keling AI web version had 17.65 million visits in September, ranking 35th in the world and ranking first in monthly visits to AI overseas.
Shengshu Technology Vidu has been online for two months. In September, the number of visits to the web version reached 11.83 million, ranking 59th in the world, and ranking 7th in the global growth list with a growth rate of 114.19%.
In addition, on September 24, ByteDance also officially announced its entry into AI video generation. Volcano Engine, a subsidiary of ByteDance, held a conference in Shenzhen and released two large models of Doubao Video Generation-PixelDance and DouBao Video Generation-Seaweed. They focus on sequential complex action commands and interactive capabilities, and are open for beta testing for the enterprise market.
On September 25, Jimeng, a Byte C-end product empowered by large video models, officially opened invitations for internal testing.
Tan Dai, President of Volcano Engine, said, "There are many difficulties in video generation that need to be overcome. The two Doubao models will continue to evolve, explore more possibilities in solving key problems, and accelerate the expansion of the creative space and application of AI videos."
On September 26, Meitu announced that its new AI short film creation tool MOKI is officially open to the public.
In addition, pika (founded in the United States by Guo Wenjing, the daughter of Xinyada’s actual controller Guo Huaqiang) has been opened to all users. Recently, Pika 1.5 also introduced a set of innovative effects called Pikaffects, which allows users to easily and significantly change video elements, such as: expansion, explosion, crushing, melting, extrusion and cake formation. However, the number of visits in September did not enter the top 100 in the world. .
According to a research report by Pacific Securities on October 20, the number of visits to the web version of domestic AI video model applications is growing rapidly and has significant advantages. Its rapid development and application are expected to empower media sectors such as advertising and marketing, film and television, and games, accelerating the cost reduction and efficiency increase in the industry and the revaluation of content assets.
Feng Cuiting, an analyst at Cinda Securities, believes that AI video goes far beyond video generation. From AI generation to AI workflow, one-stop AI video generation + editing + story creation is expected to become the core development direction of the industry.
The Open Source Securities Research Report pointed out that the continuous improvement of the multi-modal capabilities of models such as search and video generation, as well as innovative breakthroughs in AI hardware, will continue to promote the commercialization of AI applications, and it is recommended to continue to deploy AI applications.