Term
Token
A token is the smallest unit a language model works with internally — usually a sub-word fragment, occasionally a single character.
Token — explained in more detail
Language models do not process text letter by letter, nor strictly word by word, but in tokens — small chunks that typically correspond to sub-word fragments. The exact split is decided by each model’s tokenizer. A rough rule of thumb: about four characters of English text equal one token. German tends to need more tokens per word due to compound nouns and inflection.
Practical relevance
Tokens are the central billing unit for API calls: input and output tokens are usually priced separately. They also drive context-window usage and response latency — longer inputs cost not only more money but also more time.
Distinction from similar terms
A token is not the same as a word: a long word like “Versionskontrollsystem” is typically split into several tokens. It is also not to be confused with authentication tokens (API keys and the like) — same word, entirely different concept.
Discover more
AI Workflows by Keyword: How We Make Recurring Routines Enforceable
A typed keyword triggers a fixed AI routine — and every single step must be committed before the next one appears. Why that's the actual trick.
GlossaryAI Watermarking
A method AI providers use to mark generated content so machines can recognise it. In text this is an invisible statistical pattern in word choice; in files it is signed provenance metadata. The goal is evidence that a model was involved in producing the content.
EncyclopediaLLM Hallucinations — Causes and Remedies
Why LLMs confidently invent falsehoods, what types of hallucinations exist, and which remedies actually help — RAG, source enforcement, verification.