Phtoo: VCG

Kimi K3, a large artificial intelligence (AI) model developed by Chinese company Moonshot AI, was launched on Friday, marking a new milestone in open-source AI development and highlighting efforts to foster more inclusive and collaborative innovation.

With 2.8 trillion parameters, Kimi K3 has become the world's largest open-source AI model to date, according to Moonshot AI. The model has shown strong performance across multiple evaluations, placing it among the leading AI systems globally.

Within hours of its release, Kimi K3 topped the Arena rankings for AI coding tools, becoming the first Chinese AI model to reach the No.1 position on the influential platform.

Industry observers said the achievement reflects not only technological progress, but also the growing role of openness, collaboration and ecosystem building in shaping the future of global AI development.

Open source: a pathway to AI innovation

As global competition in AI intensifies, the focus has expanded beyond computing power and algorithms to include openness, accessibility and the ability to build sustainable innovation ecosystems.

Moonshot AI said Kimi K3 is the first open-source model to reach the 2.8-trillion-parameter scale. Over the past year, Kimi models have remained among the largest open-source AI models for nine months.

"More parameters mean a higher ceiling for capabilities and can deliver more intelligent performance," the company said.

Experts noted that open-source development has become an important driver of AI innovation by allowing developers, enterprises and research institutions worldwide to participate in technological advancement.

"The vitality of technology lies in openness and circulation," said Liu Tieyan, president of the Zhongguancun Academy. "A number of Chinese open-source AI models are moving from individual breakthroughs toward collective progress, providing new ideas and approaches for global AI development."

The growth of open-source models reflects a broader shift in the AI industry. Rather than relying solely on closed technological ecosystems, open platforms enable wider participation, accelerate experimentation and help expand AI applications across industries.

Toward a more inclusive AI ecosystem

Beyond its technical performance, Kimi K3 highlights the broader potential of open-source AI to make advanced technologies more accessible.

By sharing technological achievements and lowering barriers to entry, open-source models can help narrow the digital divide and allow more countries, organizations and developers to participate in the AI transformation.

Some international venture capital institutions have reported growing global interest in Chinese open-source AI models. At one institution, around 80% of companies adopting open-source AI solutions reportedly used Chinese AI models.

Meanwhile, open-source AI models have accumulated more than 10 billion downloads worldwide, reflecting rising demand for accessible and adaptable AI technologies.

Data also shows growing international adoption of Chinese AI products. Earlier this year, Chinese models surpassed US competitors for the first time in token traffic on OpenRouter, a global AI model access platform. In the first week of July, Chinese models accounted for 63% of related traffic, with DeepSeek becoming one of the preferred choices among US enterprises.

Through open collaboration, countries can explore ways to ensure AI technologies contribute to economic growth, social progress and human well-being.

Promoting global AI cooperation through applications

China has continued to advocate international cooperation in AI development, emphasizing openness, inclusiveness and responsible innovation.

On Friday, the National Development and Reform Commission and relevant departments released the Action Plan for Artificial Intelligence Cooperation and Development and the case collection China Intelligence, Global Benefits (2026).

Published for the third consecutive year, the collection highlights 10 examples of international AI cooperation.

Compared with previous editions, this year's cases cover a wider range of countries and regions, showcase more diverse applications under the "AI+" initiative, and place greater emphasis on the principle of "people-centered and AI for good."

The examples demonstrate how AI technologies can support economic development, public services and social progress, reflecting efforts to share AI opportunities with the international community.

China has repeatedly called for the establishment of a fairer and more inclusive global AI governance system, advocating that AI should become a technology that benefits all humanity rather than one accessible only to a few.