The smallest chunk of text a language model reads and writes, often a word or part of one; usage and cost are usually measured in tokens.
Language models do not read text letter by letter or exactly word by word. They split it into small pieces called tokens. A token might be a short whole word, a fragment of a longer word or a punctuation mark. Depending on the model, Arabic text often needs more tokens than English to express the same meaning.
Tokens matter because most model providers measure usage and cost by the number of input and output tokens, and every model has a maximum number of tokens it can handle at once. Longer prompts and longer answers mean higher cost and more waiting time for your users.
Practical tip: keep the instructions your app repeats on every request short and clear, and do not send whole documents when a relevant section is enough. Set a sensible answer length, such as answer in three sentences, to save cost and speed up responses.