China now dominates AI video generation, with nine of the top ten models coming from Chinese companies, according to Artificial Analysis. This shift marks a crucial moment in the global AI race, where the US leads in text and code but lags in video.
The rise of models like Minimax H3, ByteDance's Seedance 2.0, Alibaba's Habby Horse, Kuaishou's Kling, and Skyworks' Skyreels V4 has propelled Chinese firms to the forefront. Meanwhile, US companies like OpenAI have seemingly deprioritized their video efforts, leaving a gap that China has swiftly filled.
The AI Video Generation Landscape: China's Rise to Dominance
Artificial Analysis, a respected model evaluation firm, recently ranked the best AI video generation models. The results are staggering: nine of the top ten are Chinese. Only Google's Gemini Omni Flash managed to break the Chinese monopoly, sitting ahead of all others.

This dominance is not accidental. Chinese tech giants have invested heavily in generative video, recognizing its potential for entertainment, advertising, and beyond. The quality of these models has improved dramatically, with outputs that are increasingly indistinguishable from reality.

Minimax H3, in particular, has gained viral attention for its spectacular results. Its ability to generate coherent, high-resolution video from text prompts has set a new benchmark. ByteDance's Seedance 2.0 and Kuaishou's Kling are also pushing boundaries, each with unique strengths in motion handling and visual fidelity.
Why China Is Winning the AI Video Race
Several factors contribute to China's lead in AI video generation. First, there is strong investment from both the private sector and government initiatives aimed at advancing AI technology. Companies like Alibaba and ByteDance have dedicated significant resources to research and development.
Second, the vast amount of data available in China, combined with a large user base for platforms like TikTok (owned by ByteDance) and Kuaishou, provides ample training material and real-world testing grounds. This data advantage accelerates model improvement.
Third, the regulatory environment in China, while strict, has allowed for rapid experimentation in generative AI, whereas in the US, concerns over copyright and deepfakes have led to more cautious approaches. This has allowed Chinese firms to iterate faster.
The US Response: A Temporary Retreat?
While Chinese firms surge ahead, US companies appear to be pulling back from AI video generation. OpenAI, which made waves with its Sora model, reportedly shelved further development in March 2026. This decision came despite Sora's initial success and hype.
The reasons for this retreat are multifaceted. High computational costs, ethical concerns, and a strategic focus on other AI areas like text and code may have contributed. However, this pause could be temporary. US firms may be regrouping to launch more advanced models in the future.
Google, with its Gemini Omni Flash, remains the sole non-Chinese model in the top tier. Its presence indicates that US companies still have the capability to compete, but they must prioritize video generation to avoid being left behind.
Key Takeaways
- Nine of the top ten AI video generation models are Chinese, per Artificial Analysis.
- Minimax H3, Seedance 2.0, Habby Horse, Kling, and Skyreels V4 lead the pack.
- Google's Gemini Omni Flash is the only non-Chinese model in the top ten.
- OpenAI shelved its Sora model in March 2026, signaling a US pullback.
- China's data advantage and investment drive its dominance.
Comparing Top AI Video Models
| Model | Company | Country | Key Feature |
|---|---|---|---|
| Gemini Omni Flash | USA | Top-ranked overall | |
| Minimax H3 | Minimax | China | Viral, high-quality output |
| Seedance 2.0 | ByteDance | China | Advanced motion handling |
| Habby Horse | Alibaba | China | Strong visual fidelity |
| Kling | Kuaishou | China | Realistic video generation |
| Skyreels V4 | Skyworks | China | Efficient rendering |
What This Means for the AI Industry
The Chinese dominance in AI video generation has significant implications. For businesses, it means access to cutting-edge video tools from Chinese providers, but also potential geopolitical and security considerations. For consumers, it promises more realistic and creative video content. The background to this story is set out in Fable 5.1: US Leads AI Frontier But China Looms. A related thread runs through Michigan Primary 2024: Democrats Face Crucial Test of. Earlier coverage in Michigan Primary 2024: Democrats Face Crucial Test for Senate shows how the argument developed.
For the US, the challenge is clear: to regain competitiveness, companies must invest in video generation research and address the ethical and regulatory hurdles. The AI race is far from over, and video is just one battleground. A closer look at the same subject appears in Apple's China-Specific AI Model: A Dual Strategy for iPhone. Anyone following the wider picture can start with AI Job Replacement in China: Real Stories and Impacts. The same theme surfaces in Ai Video: latest articles, which takes a different angle.
As the technology evolves, we can expect further breakthroughs. The competition between US and Chinese firms will likely intensify, driving innovation and lowering costs. Ultimately, users worldwide will benefit from better AI video tools. Context on how this played out before sits in Australia's Economy Limps, RBA Should Pause Rate Hikes. A companion piece, AI Debate Divides Silicon Valley as China Gains Ground, approaches the same subject from another angle.
FAQ
What is Minimax H3?
Minimax H3 is a generative AI video model from Chinese company Minimax. It has gained popularity for its spectacular, high-quality video outputs and is ranked among the top models by Artificial Analysis.
Why are Chinese AI video models dominating?
Chinese firms benefit from massive data resources, strong investment, and a regulatory environment that allows rapid experimentation. This has enabled them to outpace US companies in video generation.
Did OpenAI really shelve Sora?
According to reports, OpenAI deprioritized its Sora video model in March 2026. This aligns with a broader US pullback from AI video generation, though it may be temporary.
Which is the best AI video model?
Artificial Analysis ranks Google's Gemini Omni Flash as the top model overall, followed closely by Chinese models like Minimax H3 and Seedance 2.0.