KEY POINTS
- China’s AI models now account for 30% of global usage, up from 1.2% in late 2024.
- Chinese-language prompts rank second globally, surpassing other languages except English in token volume.
ISLAMABAD: Chinese open-source AI models now account for nearly 30 percent of global usage, a sharp rise from 1.2 percent in late 2024, according to a report by OpenRouter and Andreessen Horowitz.
This remarkable growth, which began from a modest 1.2 percent in late 2024, highlights China’s rapid rise as a key player in the AI space.
Prominent Chinese models such as Alibaba’s Qwen, DeepSeek’s V3, and Moonshot AI’s Kimi K2 have driven much of this growth, as these systems gain traction for their efficiency and cost-effectiveness.
Chinese Models Surge in Token Volume
Chinese-language prompts have become the second most-used language globally, ranking just behind English in token volume.
The report indicates that Chinese AI models processed an average of 13 percent of the weekly token volume in 2025, with growth accelerating in the latter half of the year.
By comparison, AI models from the rest of the world averaged 13.7 percent. Despite challenges such as restrictions on access to advanced GPUs from companies like Nvidia and AMD, China has emerged as a competitive force in the development of large language models (LLMs).
Alibaba and DeepSeek Lead the Charge
China’s rapid advancement in open-source AI models has been propelled by companies like Alibaba Cloud, whose Qwen models and DeepSeek’s V3 have set a high bar for development cycles.
The quick release schedules and iterative improvements have allowed these companies to stay ahead of the curve, making their AI tools both flexible and widely adopted by users across the globe.
Chinese Language Surges in AI Usage
Chinese has now become the second most used language in AI prompts, accounting for nearly 5 percent of global requests, a significant jump from its 1.1 percent share on the internet.
This surge reflects both the increasing influence of Chinese AI models and the broader adoption of AI systems by Chinese-speaking users, marking a shift in the global AI landscape.
China’s Competitive Edge in AI Development
The rise of Chinese models, despite external challenges such as restrictions on GPU access, underscores China’s competitive edge in AI model development.
The country’s rapid iteration cycles, combined with the global demand for more efficient and cost-effective AI solutions, have positioned Chinese open-source models as strong contenders in the market.
Fragmented AI Landscape
The global market for open-source AI models has evolved from a DeepSeek-dominated monopoly in late 2024 to a fragmented landscape by late 2025.
No single model now dominates the space, with companies like Alibaba Cloud’s Qwen and Moonshot AI’s Kimi adding to the growing competition.
