China leverages short-video ecosystem, lower costs to carve out lead in AI video

Models from Alibaba, MiniMax and ByteDance currently dominate benchmark rankings on industry evaluation platform Artificial Analysis

Chinese artificial intelligence companies are carving out a distinct lead over their US rivals in video generation, leveraging vast short-video ecosystems, looser copyright rules and aggressive pricing to dominate an arena where compute alone does not dictate success, analysts say.

While US tech giants such as OpenAI and Anthropic continue to lead in frontier closed-source large language models (LLMs), Chinese firms are emerging as some of the most formidable competitors in the AI video generation battleground.

Models from Chinese firms including Alibaba Group Holding, MiniMax and ByteDance currently dominate benchmark rankings on industry evaluation platform Artificial Analysis.

Alibaba’s Wan 3.0 ranks first on Artificial Analysis’ text-to-video leaderboard with audio, ahead of second-placed Google Gemini Omni Flash. A version of MiniMax H3 post-trained by US platform fal ranks third, followed by the original open-weight H3 and ByteDance’s Seedance 2.0. Together, Chinese models occupy eight of the top 10 positions.

Chinese models also lead across several related categories, including image-to-video and video editing, based on blind user comparisons using identical prompts. Alibaba owns the South China Morning Post.

Industry analysts attribute China’s video AI momentum to a combination of proprietary training data, rapid iteration cycles, and strong domestic demand for short-form digital content.