Chinese artificial intelligence startup Moonshot AI has unveiled Kimi K3, a next-generation AI model it says can compete with leading systems from OpenAI and Anthropic, marking another significant step in China’s push to narrow the global AI gap.
The model, introduced at the World Artificial Intelligence Conference (WAIC) in Shanghai, boasts 2.8 trillion parameters, a key measure of an AI system’s scale and computational capability. Its full suite of coding, reasoning and knowledge-processing features is expected to become available when it is released as an open-source model on July 27.
If successful, Kimi K3 will become the world’s first freely downloadable open-source AI model in the nearly three-trillion-parameter class, allowing developers to modify, customise and deploy it for a wide range of applications.
The launch underscores the rapid progress Chinese AI firms are making despite US export restrictions on advanced chips and AI technologies.
For years, American companies such as OpenAI, Anthropic and Google have been viewed as the clear leaders in frontier AI. However, Moonshot AI’s latest announcement suggests Chinese developers are closing the performance gap faster than many industry observers expected.
Backed by Chinese technology giants Alibaba and Tencent, Moonshot AI said Kimi K3 is its “most capable flagship model to date.”
Unlike proprietary AI systems developed by many US companies, Kimi K3 will be open source, enabling researchers and businesses worldwide to inspect, modify and adapt the model to their specific needs.
The company said the model is designed to perform complex engineering, coding and reasoning tasks with minimal human supervision.
According to Moonshot AI, independent evaluations by Artificial Analysis and Arena.ai indicate that Kimi K3 performs on par with leading American AI models, including OpenAI’s GPT family and Anthropic’s Claude.
In one benchmark focused on web interface engineering, Kimi K3 reportedly outperformed Anthropic’s Fable model in blind human preference tests.
Although the model’s enormous size means organisations will need substantial computing resources to run it locally, analysts believe its open-source availability could accelerate AI innovation while increasing competitive pressure on commercial AI providers.
Kimi K3’s launch comes at a pivotal moment for the global AI industry.
The release follows recent scrutiny of advanced AI systems in the United States, where concerns over cybersecurity and national security have prompted tighter oversight of frontier AI technologies. Washington has increasingly classified cutting-edge AI models as strategic assets, imposing stricter controls on their development and export.
Despite those restrictions, Chinese AI companies continue to advance rapidly, developing increasingly sophisticated models using domestic expertise and investment.
