Verified65% confidenceFactExact time
Large model APIs typically use tokens as the unit of measurement; in a Chinese-language context one Chinese character is approximately 1–2 tokens, while in English one word is approximately 1–1.5 tokens.
3
Sources
65%
Confidence
Long-term
Relevance
8/6/2026
First Seen
Sources
Kimi K3为何被催更？一条Reddit吐槽揭示国产大模型竞争真相
Related Claims
VerifiedLLM APIs are billed by token, with approximately 1–2 tokens per English word and approximately 1–2 tokens per Chinese character.78% similarUnverifiedThird-party API calls are typically measured in tokens, with each English word being approximately 1–2 tokens and each Chinese character being approximately 1–2 tokens.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/694892API
curl https://kongchang.com/api/v1/knowledge/claims/694892MCP
get_claim(id=694892)