Token
What is Token?
The unit a model reads and writes. In English one token is about three-quarters of a word (roughly four characters), though this varies by language. Pricing, speed and context limits are all counted in tokens, so token count is what you actually pay for.
Why it matters
A model does not read words the way we do. It breaks text into tokens, and everything commercial is counted in them. Price, speed and the amount of text a model can handle at once are all measured in tokens, so token count is what shows up on the invoice. For anyone budgeting an AI feature this is the unit that matters. A chatbot that quietly feeds the model a long instruction and the whole conversation history on every message can cost several times more than one that keeps its inputs tight, even though the user sees the same reply.
In practice
In English a token runs about four characters, roughly three-quarters of a word, though this varies a lot by language. A finished sentence might be 20 tokens. When you estimate the running cost of an AI feature, you count the tokens going in (the instruction, the user’s text, any documents) plus the tokens coming out, then multiply by the per-token price. Cut the input and you cut the bill.
