Chinese artificial intelligence startup Moonshot AI on Monday fully released its latest flagship large language model, Kimi K3, as open source. The move marks another significant milestone in China's rapidly growing open AI model ecosystem.
Kimi K3 has 2.8 trillion parameters and is currently the largest open-source AI model in the world by parameter count.
The release includes the model's complete weights, a technical report, and three core infrastructure technologies used during training. Developers around the world can download the model, run it locally, and modify it further.
China has previously released several AI models as open source, but this is the first time a flagship model with 2.8 trillion parameters has been made fully available to the global developer community.
China's open-source AI development accelerates
The pace at which large AI models are being released in China has increased significantly this year. Several leading technology companies have made their most advanced models publicly available.
Within just six months, several Chinese developers have opened access to their most capable foundation models.
According to a report by the Hugging Face platform titled State of Open Source on Hugging Face: Spring 2026, open-source models developed in China account for 41 percent of all downloads worldwide.
The report also noted that six of the most widely used large language models in the world come from Chinese development teams. Over the past twelve months, Chinese projects held the world record for the largest open-source model for nine months, while simultaneously maintaining the fastest pace of technological innovation.
From models with one trillion parameters last year, to models with 1.6 trillion parameters, and now to the current 2.8 trillion parameters, China has repeatedly pushed the boundaries of open AI system size.
China's open-source ecosystem today spans models ranging from hundreds of millions to hundreds of billions of parameters. Smaller models can run offline on smartphones and industrial devices, while larger systems are capable of handling complex research and analytical tasks.
Users thus have solutions available for both edge devices and large-scale enterprise systems.
From open models to industrial application
Unlike closed models, where users can only access AI capabilities through a service provider, open-source models release their weights.
Developers can therefore download them, deploy them independently, and adapt them for their own applications.
Chinese open models are no longer limited to general-purpose conversational assistants. They are increasingly finding use in specific industrial sectors.
Specialized open-source models have been developed for the automotive industry, manufacturing, energy, and healthcare, among others. They embed domain expertise and knowledge directly into AI systems, helping bridge the gap between general-purpose models and their practical application.
According to official figures, the total number of downloads of Chinese open-source models has already exceeded ten billion, the highest in the world. Industry statistics show that six out of every ten downloads of large models worldwide now come from systems developed in China.
China's open AI ecosystem has, according to the report, moved from individual technological breakthroughs to a broader and more mature system.
As technologies continue to develop and barriers to deployment decrease, AI is moving out of research laboratories and into manufacturing, energy, services, and other areas of the real economy.
It is helping to modernize industries, improve efficiency, and reduce costs for businesses.
The open approach simultaneously makes it easier to deploy AI in industry. Large language models are thus transforming from general-purpose conversational tools into productive systems capable of performing specific tasks, supporting industrial operations, and increasing efficiency in manufacturing, energy, transportation, and finance.
Government support
Zhong Xinlong, a researcher at the Future Industries Research Center of the China Center for Information Industry Development, noted that in the past, the most advanced AI models were largely controlled by a small number of closed platforms.
Businesses therefore had to purchase the AI capabilities they needed as a service.
As open-source models develop, companies can build their own applications combining their data, business processes, and domain knowledge. The result is solutions better tailored to specific needs and circumstances.
China is also strengthening policy support for open AI.
China's Ministry of Industry and Information Technology, together with seven other government agencies, this year issued an implementation guideline for the "AI Plus Manufacturing" initiative.
The document identifies building advanced open-source AI communities and launching model open-source projects as key priorities.
By 2027, it aims to support the deeper use of three to five general-purpose large models in manufacturing, the creation of one hundred high-quality industrial datasets, and the expansion of AI into 500 representative industrial sectors.
Zhong said that open-source models are changing the global AI competition, which is shifting from the dominance of a few technology giants toward broader competition among individual industries.
"AI sovereignty is becoming a significant policy priority for many countries," he said. "Open-source development broadens participation and enables more languages, regions, and industries to join the AI ecosystem."
China now holds 60 percent of the world's AI patents, the highest share globally. According to estimates from the Ministry of Industry and Information Technology, the value of China's core AI industry exceeded 1.2 trillion yuan in 2025, with more than 6,200 AI-focused companies operating in the country.
As China's open-source ecosystem continues to expand, AI is expected to be integrated ever more deeply into the real economy.
It should thus create long-term momentum for new industrialization and high-quality development of the digital economy.
gnews.cz/CMG
