On the 28th, the reporter learned from the State Administration for Market Regulation that the "Artificial Intelligence Big Model" series of national standards has recently been officially implemented, marking a new stage of "scientific authority and unified standardization" for China's big model industry. The relevant person in charge of the State Administration for Market Regulation introduced that as the first national standard focusing on general large models, this series of standards fills the gap in the technical evaluation system, clarifies the requirements for performance, safety, and service capabilities, and the supporting evaluation capabilities have been recognized by the China National Accreditation Service for Conformity Assessment. Based on this, the "Qiusuo" - LMBench evaluation benchmark integrates methods, datasets, and automation platforms, becoming an industry recognized "capability ruler". The evaluation system has formed a mainstream big model whitelist, providing a basis for the selection of central and state-owned enterprises, and supporting the construction industry big models such as Sinopec and Southern Power Grid. It has completed scenario verification of models such as Huawei Pangu, iFlytek Spark, and China Mobile's "Nine Heavens". As of now, the standard tool has completed over a thousand evaluation tasks, called the large model over 950000 times, accurately identified common issues such as illusion control and content security, assisted nearly 30 manufacturers in technological iteration, and promoted the formation of a "research and development evaluation application upgrade" closed loop. In addition, the internationalization of standards is also being promoted simultaneously. Developed English, Laotian, and Khmer versions for ASEAN, filling regional gaps; The Chinese open source model architecture has been adopted by projects in countries such as Singapore, demonstrating its global influence under the guidance of standards. (New Society)