According to the latest OpenRouter data, in the week of October 12–18 global AI large-model calls totaled 164 trillion tokens, down 4.65% week on week. Among the listed models, Chinese large models accounted for 79.18 trillion tokens (down 8.72%), while US models were 17.36 trillion (down 8.49%). China's weekly call volume has exceeded the US for 25 consecutive weeks, holding the global top spot, with the global top five by weekly calls all being Chinese AI models.
The data reinforces China's lead in call scale, though the sequential decline on both sides hints at demand volatility that should not be ignored. The industry trend is shifting from "parameter competition" to "calls, cost and scenario competition," which may favor domestic compute, inference services and AI applications in the near term.
In related standardization news, the State Administration for Market Regulation recently released three national guidance documents for AI in the steel industry — a scenario-classification guide, a dataset technical requirement, and a service-capability maturity assessment — building a complete closed loop of "scenario classification – dataset norms – maturity assessment" to support enterprises' digital-intelligence transformation and move steel-industry AI from pilot to scaled deployment.
Also on October 19, nine departments including the Ministry of Commerce issued opinions on accelerating "technology-prospering commerce," proposing to strengthen R&D and commercialization, cultivate commercial technology-application innovation centers, and accelerate the market conversion of frontier technologies such as brain-computer interfaces and embodied intelligence, while supporting smart warehousing and digital supply chains.
Source: Daily Economic News / Tencent News
