Chinese large language models have dominated global token usage for 15 consecutive weeks, with models from China now accounting for 34.25 trillion weekly tokens. The DeepSeek-V4 Flash model has achieved the top position, experiencing a 570% week-on-week increase in usage. This surge places Chinese models significantly ahead of their US counterparts, which collectively use 9.17 trillion tokens. AI
IMPACT Confirms the growing dominance of Chinese LLMs in global token usage and highlights DeepSeek-V4 Flash as a leading model.
RANK_REASON Frontier lab model release with significant usage metrics. [lever_c_demoted from frontier_release: ic=2 ai=1.0]
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
