Verified70% confidenceConceptExact time
Large language model APIs usually charge by the number of input and output tokens; a token is the smallest unit of text processed by the model, roughly equal to one English word or half a Chinese character.
4
Sources
70%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
Unverified大语言模型以词元（token）为基本处理单位，而非以单词为单位理解语言75% similarUnverifiedAPIs are billed per token; one token corresponds to approximately 3/4 of an English word or 1–2 Chinese characters.74% similarUnverified大语言模型的用量计费通常基于token（词元）消耗，而非简单的对话次数74% similarUnverified大语言模型以token为单位处理文字而非以字母为单位，这解释了早期模型在字母计数任务上出错72% similarUnverifiedA token is the smallest unit by which large language models process text, roughly corresponding to half to one English word or half a Chinese character.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1024644API
curl https://kongchang.com/api/v1/knowledge/claims/1024644MCP
get_claim(id=1024644)