China has reached a new stage in the global race for dominance in artificial intelligence, after the startup "Moonshot AI," backed by Chinese technology and e-commerce empire Alibaba Holdings, unveiled a new AI model that it says competes with the best models of the American companies OpenAI and Anthropic.

The launch of the new Chinese AI model caused a stir in the sector, fueling fears of erosion of the superiority of American companies over Chinese companies in this critical field.

According to the Chinese company, the new AI model "Kimi K3," which consists of 2.8 trillion parameters, is the largest open-source AI model in the world so far.

It is noted that the number of parameters is an indicator of the capability and complexity of any system. The company will make this model fully available and free to developers in all countries of the world by the end of July this year.

Related News

Moonshot said while announcing the model's test results that "Kimi K3" outperforms comparable American AI models such as Claude Opus 4.8 from Anthropic and GPT-5.5 from OpenAI, at least in performing coding tasks and in complex applications.

At the same time, Moonshot acknowledged that the new model is still slightly less efficient than leading American models, especially Claude 5.5 from Anthropic and the brand new GPT-5.6 Sol from OpenAI, but it noted that the gap with superior American models is rapidly narrowing.

Observers say this development is reminiscent of the "DeepSeek moment" in early 2025, when the emergence of the Chinese AI model DeepSeek challenged perceptions of American dominance, leading to a sharp decline in shares of global technology companies.

Since then, the importance of Chinese AI models has grown globally, partly because many of them are available as free open-source software, making them generally cheaper to run than their Western counterparts.

Related News