AI Glossary
Token
The basic unit of text that AI models process. A token is roughly 3/4 of a word. Understanding tokens helps you estimate API costs and stay within model context limits.
Understanding Token
Tokens are how AI models see text. The word "understanding" might be split into "under" and "standing" as two tokens. On average, 1 token equals about 0.75 words, or 4 characters in English.
Tokens matter for two practical reasons: cost and context limits. AI APIs charge per token processed (both input and output), so understanding token counts helps you estimate costs. Context windows are measured in tokens, determining how much text you can process in a single interaction.
For business planning, here are useful benchmarks: a one-page document is roughly 500 tokens, a typical email is 100-300 tokens, and a 50-page report is about 25,000 tokens. At current GPT-4o pricing, processing that 50-page report costs roughly $0.06.
Token in Canada
French text typically uses 10-20% more tokens than equivalent English text due to longer words and grammatical structures, which slightly increases API costs for bilingual Canadian applications.
Related Services
Frequently Asked Questions
Roughly divide the word count by 0.75 (or multiply by 1.33). A 1,000-word document uses approximately 1,333 tokens. OpenAI's tokenizer tool provides exact counts for precise cost estimation.
Tokens are the actual units the model processes. Different words consume different numbers of tokens, and tokens capture sub-word units that are more computationally meaningful than whole words.
See Token in Action
Book a free 30-minute strategy call. We'll show you how token can drive real results for your business.