DeepSeek V4 Flash ranked first in OpenRouter’s weekly model-usage ranking for July 27 to Aug. 2, with 7.22 trillion tokens processed on the multi-model aggregation platform. On Aug. 1, the model processed 8 trillion tokens on OpenCode, including 5 trillion tokens consumed through free trials and 3 trillion paid by developers.
V4 Flash is a high-throughput model aimed at developers and business users. Chinese models occupied the top four positions on OpenRouter’s weekly ranking, while DeepSeek’s V4 Flash 0731 and V4 Pro also ranked in the top six. The figures measure activity on the tracked platforms rather than DeepSeek’s total global usage. [Jiemian, in Chinese]