Real World Maths Busters

The Maths Behind Everyday Life

The story, the formula and a worked example — see exactly where this maths shows up in the world around you.

Back to all stories
Technology 3 min read

What an AI's "context window" really holds

Tokens are chunks of words — roughly four tokens for every three words of English.

AI models measure memory in tokens, not words. A token is a word-piece, and English runs about four tokens for every three words, so a context window converts into a readable page count you can plan around.

Tokens from words

Multiply the word count by about 4/3 to estimate tokens. Divide a context window by that ratio to get words it can hold.

tokens≈words×43\text{tokens} \approx \text{words} \times \dfrac{4}{3}

Rough English conversion

A 1500-word document

Estimate the tokens.

Words to tokens

  1. 1

    Multiply

    1500×43=2,0001500 \times \frac{4}{3} = 2,000
  2. 2

    Flip it

    A 128,000-token window holds roughly 96,000 words — a short novel.

Answer

About 2,000 tokens — handy for knowing what fits in a prompt.

The formula behind it

Tokens

words×4/3\text{words} \times 4/3

The takeaway

Tokens are word-pieces, about 4 per 3 English words. Converting lets you judge how much text fits an AI context window.

Back to all stories