TEL:0571-83868908
About Souei
Now:HomeAbout SoueiIndustry News

China's large-scale AI model uses over 60% Chinese data for training.

People's Daily Japanese Edition 19 August 2025 15:29

According to the National Data Bureau, Chinese data plays a crucial role in the improvement of training performance for large AI (artificial intelligence) models in China. Over 60% of the data used for training most of China's large AI models consists of Chinese data, with some models reaching up to 80%. The development and supply capabilities of high-quality Chinese data continue to improve, driving the rapid performance enhancement of China's AI models, as reported by the People's Daily.


In the AI era, "tokens" refer to the smallest unit of text processing. Liu Liehong, the director of the bureau, explained that "at the beginning of 2024, China's daily token consumption was 100 billion, but by the end of June this year, it had surpassed 30 trillion, an increase of over 300 times in just a year and a half. This highlights the rapid growth of AI applications in China." (Edited by KS) "


People's Daily Online Japanese Version" 19 August 2025.