The China-US AI gap has fallen to around 3%, the lowest on record, according to a Bloomberg Intelligence report. Chinese artificial intelligence models have drastically reduced the performance difference relative to American leaders.
What happened?
According to Bloomberg Intelligence analysis, the China-US AI gap dropped from about 9% in May and 15% earlier in the year after the September release of DeepSeek’s V4.1 Flash model. Senior analyst Robert Lea notes that the advance raises questions about how long the United States can maintain its technological lead and could help Chinese companies gain a larger share of the global market.
Besides DeepSeek, models from Alibaba (Qwen), Z.ai (GLM), Moonshot (Kimi) and even Xiaomi appear among the most used on platforms such as OpenRouter. Zhang Yaqin, founding dean of Tsinghua University’s Institute for AI Industry Research, said that before DeepSeek China was two to three years behind; today the lag is roughly one model generation (two to three months).
Why does it matter?
The narrowing of the China-US AI gap shows how fast China’s open-weight and low-cost model ecosystem is advancing. While Western labs focus on closed high-performance models, China multiplies efficient options that already lead in usage volume on some platforms.
- Current gap: ~3% (record)
- Gap in May: ~9%
- Main Chinese players: DeepSeek, Qwen (Alibaba), GLM (Z.ai), Kimi (Moonshot), Xiaomi
- Source: Bloomberg Intelligence / China Daily
What changes in practice?
Developers and companies gain more competitive Chinese model alternatives, often cheaper and with open weights. Global competition for AI capability becomes even fiercer, and China positions itself as a relevant force in both research and practical adoption.
Image credit: China Daily / Bloomberg — Source: China Daily and Bloomberg Intelligence
By GeekikiBot
