[DeepSeek's Large-Scale Model Ranks First Globally in Weekly Calls] According to Jinse Finance, on October 26th, based on the latest data from OpenRouter, the total global AI large-scale model calls last week (October 19th to 25th) reached 188 trillion tokens, a 14.63% increase week-on-week. Among the listed AI large-scale models, Chinese AI large-scale models saw a weekly call volume of 93.54 trillion tokens, an 18.14% increase week-on-week; while US AI large-scale models saw a weekly call volume of 22.47 trillion tokens, a 29.44% increase week-on-week. Chinese large-scale models have surpassed the US in weekly call volume for 26 consecutive weeks, firmly holding the top spot globally. The Daily Economic News noted that last week, the top five global AI large-scale models in terms of call volume were all Chinese. DeepSeek V4.1-Flash-1017 ranked first, with a weekly call volume of 33.6 trillion tokens, an 801% increase week-on-week. DeepSeek V4.1-Flash-1017 was officially updated on October 17th. This update, based on the V4.1 Flash released on September 10th, improves performance in inference, programming, automation, and vision tasks. DeepSeek claims the new model has achieved significant improvements in various benchmarks and has reached new Pareto optimal performance-cost across multiple Harnesses (control frameworks), representing the best cost-performance ratio. DeepSeek V4.1 Flash ranked second with weekly call volume of 18 trillion tokens, a 45% decrease compared to the previous week; MiniMax M3.1 Flash (free) surged to third place with weekly call volume of 13.7 trillion tokens. (Daily Economic News)