China’s artificial intelligence (AI) model is rapidly eroding the share of U.S. companies such as Anthropic’s Claude and Google’s Gemini in the global coding market. It is expanding its influence by introducing a series of open-source products that are comparable to U.S. Frontiers in terms of performance as well as price competitiveness, which was considered a strength of existing Chinese models. In particular, the pace of expanding its presence in emerging markets such as the Middle East and South America is remarkable. Although new models are steadily being born, it is compared to Korea, which has little presence.
According to the information technology (IT) industry on the 7th, the global share of Claude and Gemini in the programming sector has steadily declined, while China’s AI has risen significantly. According to OpenRouter, as of August 11 compared to July 21, the Anthropic Claude SONET 4 share fell 15.7 percentage points in the programming area, recording the biggest drop. The Gemini 2.5 Pro and Flash also decreased by 3.6 percentage points and 4.4 percentage points.
On the other hand, Alibaba’s Qwen 3 coder grew by 16.4 percentage points during the same period, accounting for 21.5 percent of the market share. In particular, Qwen’s growth was remarkable. While DeepSeek is slowing down, it has also ranked first among Chinese models in terms of performance. Alibaba’s frontier model “Qwen3-235B-A22B-Thinking-2507” scored 64 points, ahead of DeepSeek’s latest model V3.1 (60 points), according to Artificial Analytics indicators.
Chinese start-ups are also chasing after them. Z, which was released in July, according to the same survey by OpenRouter.AI’s GLM 4.5 and Moonshot AI’s Kimi-K2 had market share of 6.1% and 3.2%, respectively, as of August 11.
Among them, Kimi-K2 attracted so much attention that it was evaluated as bringing another “deep moment.” An industry official said, “Now, most of China’s open source models, as well as DeepSeek, have competitive edge to compete with U.S. big tech models in terms of functionality beyond cost-effectiveness.”
Although there are many performances, the biggest reason why Chinese models stand out in the global market is their price competitiveness. The Qwen 3 coder costs $1 per 1 million token of input and $5 per 1 million token of output, which is cheaper than the Claude Opus 4 (input $15, output $75). The startup model is more unconventional. Z.AI’s GLM 4.5 is $0.6 per 1 million tokens input and $2.2 output, the lowest among Chinese models. MoonshotAI’s Kimi-K2 is $0.6 input and $2.5 output.
This price competitiveness is particularly strong in emerging countries such as the Middle East and South America than in the United States or Northeast Asia. Qwen and Z.Analysts say that price competitiveness is effective against the background of Chinese models such as AI showing rapid growth in the coding market in emerging countries. Similar web data showed that Qwen models, excluding China, accounted for 27.5% of traffic in Iraq, 19.1% in Brazil and 12.1% in Turkiye. Z.AI also has offices in the Middle East and Africa to supply AI solutions to local governments and state-owned companies.
The rise of China’s open-source model is not just a corporate-level result. Since the “deep shock” earlier this year, China has established an open-source strategy nationwide and provided full support to related companies. This year, Chinese companies launched a series of frontier models and their global share rose due to the government’s support.
As such, the Chinese model has emerged rapidly this year and is competing in the U.S. and global coding markets, while the Korean model’s presence is still insignificant. Although LG AI Research Institute’s recently released ‘Exemployee 4.0’ was evaluated as being at the top of the global rankings in coding performance, it is far from the actual market share. The reality is that many Korean companies use overseas models. Industry experts point out that it is urgent to strengthen the coding sector’s capabilities, one of the key areas of AI competitiveness.
[Reporter Ahn Seonje]