Photo: Courtesy of Moonshot AI

As Chinese AI startup Moonshot AI’s newly released model Kimi K3 gained wide attention, the company on Monday released its model weights and technical report, and open-sourced the key infrastructure technologies that supported its training, in a bid to accelerate the deployment and popularization of frontier intelligence and promote AGI research.

According to the company, at present, everyone can download and deploy the Kimi K3 model — whether for internal R&D or embedding it into end-user products, it can be used freely.

The company noted that it firmly believed in the value of open-weight models, which lower the barrier to accessing intelligence, drive innovation, and empower users with greater control over their data, as well as stronger privacy protection and ownership.  

“We believe that for a technology as far-reaching as AGI, a broad and open ecosystem is the best soil to nurture it. To this end, we will continue to contribute our part,” Moonshot added.

Moonshot described Kimi K3 as the world’s first open-source model at the 3-trillion-parameter level, designed for advanced intelligent scenarios including long-horizon coding, knowledge work and reasoning tasks. It is a 2.8 trillion parameter Mixture-of-Experts model with native visual understanding capabilities and support for a 1 million token context window. 

Since its release on June 16, the Kimi K3 has received greater visits than expected, and it decided to suspend new C-end subscriptions.

Bloomberg reported on Monday that Moonshot developed K3 with 2.8 trillion parameters — a rough measure of AI complexity — making it the world's largest open-weight model at a time when leading US AI companies like Anthropic are keeping the size of models like Claude private.

In stark contrast to the US’ closed-source model, China has pursued an open-source approach, which has sparked positive reactions from many countries and regions around the world. 

Meanwhile, some US officials are reportedly calling for investigating and threatening to sanction Chinese AI firms. In response, a spokesperson for the Chinese Ministry of Commerce said on Monday that such moves lack factual basis and legal justification, and that China will take all necessary measures to firmly safeguard its legitimate rights and interests.  

Innovation is not the exclusive domain of any single country. Chinese AI companies have long been committed to advancing fundamental research, balancing technological self-reliance with open cooperation. Guided by the principles of consultation, joint contribution and shared benefits, they have driven the rapid development of AI technologies and achieved notable progress in areas such as coding capabilities, the spokesperson said.

Global Times