Chinese open-source AI models gain global market share

by Alimat Aliyeva

In November last year, Chinese AI models significantly increased their presence in the global market, rising from 1% to 15%, thanks in large part to the widespread use of open-source code, according to data from Nikkei, Azernews reports.

Statistics show that over 40% of AI models developed by Chinese companies are applied to complex tasks, including software development. Alibaba's Qwen continues to be the world's most popular open-source AI platform, with more than 700 million downloads as of this month. Alibaba offers a wide range of open-source AI models, with parameter sizes ranging from 600 million to tens of billions, catering to diverse business and research needs.

Last year’s success of DeepSeek’s first AI model drew global attention to Chinese AI innovations, particularly because many Chinese developers keep their source code open. This openness allows third-party developers to adapt the software for a wide variety of applications, contributing to the rapid rise in international popularity of Chinese AI solutions.

Currently, DeepSeek is preparing to launch its next-generation AI model. In the latest Nikkei ranking, DeepSeek’s model released in December achieved the ninth position out of 92 for Japanese language performance. Among open-source AI models, DeepSeek offers the fastest performance, followed by Alibaba Qwen, while Google’s and OpenAI’s open-source models lag behind both. In Japan, six out of ten AI models developed by local companies are based on DeepSeek and Qwen technologies.

These developments highlight a notable shift in the AI landscape: Chinese open-source solutions are not only competing but often outperforming established global players, reshaping how AI tools are developed, shared, and applied around the world.