Token

A token is a text building block a language model processes — usually a word, a syllable or a punctuation mark. As a rule of thumb, 1,000 tokens correspond to about 750 words. Models compute and ‘think’ in tokens, not words.

For economics the token is the decisive measure: cloud providers bill usage-based per 1M tokens, split into input and output. Low-cost models sit in the range of a few cents, powerful ones several times higher.

This billing makes success expensive: the more a use case is used, the higher the token bill. When you run it yourself this usage-dependent curve disappears — there you pay for capacity, not per token.