DeepSeek's Latest Flash Model Dominates Global LLM Rankings with 801% Weekly Growth

DeepSeek's V4.1-Flash-1017 model surged to the top of global API usage rankings this week, processing 33.6 trillion tokens in just seven days—an extraordinary 801 percent jump from the prior week and cementing Chinese AI development's growing dominance in the commercial LLM market. The October 17 release of the optimized variant triggered a rapid migration from the earlier V4.1-Flash, which saw usage drop 45 percent to 18 trillion tokens, as developers shifted to the newer version within days.