Chinese AI developers are increasingly embracing open source to speed up technological iteration and ecosystem development.

Recently, MiniMax officially open-sourced its H3 model. Within just three days, it had topped the trending chart on Hugging Face, the world's largest open-source AI platform. Over a hundred enterprises integrated it immediately, with full-stack compatibility from chip makers to inference frameworks.

MiniMax H3 showcased its dominance in video generation, ranking first across three categories — multi-image-to-video, image-to-video, and video editing — on the Design Arena platform.

Emad Mostaque, an iconic figure in the open-source community and founder of Stable Diffusion, said on X, "Bravo to MiniMax AI some interesting things in here."

This is not just a case of overnight success for a single model, but further solid evidence of the collective rise of China's open-source AI models. Major players like DeepSeek, Kimi, and Qwen have released and open-sourced their latest-generation models.

The primary value of open source lies in technological spillover. It tears down barriers of closed-door R&D, making advanced model capabilities accessible to a broad range of developers and enterprises instead of being available exclusively to a handful of tech giants. Such spillover quickly energizes the entire industrial chain: Chip vendors can optimize hardware tailored to model features, inference frameworks deliver instant compatibility, and application developers pursue innovation at far lower costs. The full-stack compatibility rolled out immediately after MiniMax H3's launch exactly illustrates such synergy.

Furthermore, open source is reshaping how technology evolves. Improvements contributed by diverse teams flow back into the model ecosystem at an accelerated pace, creating a self-reinforcing cycle of continuous advancement. This mechanism boosts efficiency and strengthens the overall competitiveness of China's AI system. Open source is not only a technical choice for Chinese companies but also a competitive strategy to expand market space through ecosystem building and openness.

The significance of this strategy stretches far beyond individual companies or a single country. For global developers, Chinese open-source models deliver more cost-effective alternatives and significantly lower the barriers to deploying AI.

According to The Economist, Martin Casado, a partner at the U.S. venture-capital firm Andreessen Horowitz (a16z), revealed that up to 80 percent of U.S. AI startups use Chinese open-source models during their fundraising pitches. MiniMax now serves over one million enterprises and developers worldwide, a fivefold increase from six months ago.

According to Hugging Face, China surpassed the U.S. in monthly downloads and overall downloads, with its open-source models accounting for 41 percent of downloads between February 2025 and February 2026, compared with 36.5 percent in the U.S. Chinese large language models took the top six spots in global usage rankings during the week from March 30 to April 5, according to OpenRouter.

Industry leaders also support the pathway of open source. Jensen Huang, founder of NVIDIA, has repeatedly stated that open-source models enhance cybersecurity, accelerate innovation and knowledge sharing, and help all nations meet their development goals. He has stressed that China is poised to deliver outstanding AI technologies, and the world should keep learning from and cooperating with China.

As tributes to China's open-source innovations evolve from individuals' praise into a widespread consensus, the global benefits brought by Chinese open-source AI are vividly demonstrated.