On 25 March, according to the daily newspaper People ' s daily report yesterday, the National Data Agency was informed that:
Early 2024, Chinese daily averageCharacter(Token) Called 100 billion; by the end of 2025, jumped to 10 trillion; in March this year, it had surpassed 140 trillion, and increased by more than 1,000 times in two years。
The terminologies are understood to be the smallest information unit of a large model for processing information with measurable, priceable and tradable features of an intelligent age。
According to the report, the significant increase in the country's daily linguistic transfers also shows that, as the market-based configuration of data elements advances, a supply system of high-quality data of artificial intelligence is being developed, and a virtuous cycle of "data supply-value release" is emerging。
Recently, APPSO also wrote that in the NLP academic literature, Token, the smallest processing unit of the text, has been translated as a term:
word, which refers to the language properties; dollars, which refers to the smallest particles. Two words, one pipe, one layer。
According to FIA, at the previous annual meeting of the China Development High-Level Forum in 2026, the Director of the National Data Agency, Liu Lihoon, stated that Token 's words' were not only a value anchor in the age of intelligence, but also a "settlement unit" linking the supply of technology to business needs, offering quantifiable possibilities for the landing of business models; they also gave Token 's words' an official translation。
Another "old acquaintance" of AI - Prompt, APPSO - also made its own new proposal for translation: "The Statutory Order"。
According to APPSO, when compared to what is now known as the "tribe," the "council" can show that Prompt's vehicle is always text and that Prompt's function is to order a driver model。