← AI Engineering Roadmap

Step 3 of 15

What Are Tokens? (The Lego Bricks of AI)

What Are Tokens? — AI Engineering Roadmap by Haroon Muhammad
TL;DR: AI models don't read whole words — they read tokens, which are little chunks of text (roughly ¾ of a word each). Tokens decide how much you pay and how much the model can read at once.

Explain it like I'm 5

Think of Lego. You don't build a castle with one giant block — you use lots of small bricks. AI reads text the same way. It chops your sentence into small bricks called tokens, then builds its answer brick by brick.

What exactly is a token?

A token is a small piece of text. It can be a whole word, part of a word, or even just a space or comma. As a rough rule, 1 token ≈ 4 letters, and 100 tokens ≈ 75 words.

The word "unbelievable" might be split into "un", "believ", and "able" — three tokens. Short common words like "cat" are usually one token.

Why should you care about tokens?

Two big reasons:

A simple way to picture it

Imagine paying for a taxi by the meter. Every word you speak and every word the driver says back adds to the fare. Tokens are that meter for AI. Short, clear prompts keep the meter low.

Quick tips to save tokens

Frequently asked questions

How many tokens is a page of text?

One page (about 500 words) is roughly 650–700 tokens. A quick shortcut: multiply your word count by about 1.3.

Do I pay for both my question and the answer?

Yes. Almost every AI provider charges for input tokens (your prompt) and output tokens (the reply), usually at different prices.

What happens if I go over the token limit?

The model will either cut off older text, drop part of your input, or return an error. This is why managing context matters.

Want to know exactly where you stand on the road from using AI to building it? Take the free 3-minute AI Builder Scorecard and get your single biggest gap.