What is a token in AI and why does it matter for how much I pay?

A token is a chunk of text that an AI model reads or writes, and it matters for cost because most AI tools charge based on how many tokens you use. Think of tokens as puzzle pieces. The word "apple" is one token, but "unbelievable" might be two or three tokens because the model breaks longer words into smaller parts. On average, one token equals about three-quarters of an English word, so a 100-word paragraph is roughly 133 tokens. Why should you care? Because when you use a tool like ChatGPT or Claude, you're paying for input tokens (what you type) and output tokens (what the AI writes back). A 2025 OpenAI pricing update showed that GPT-4o costs $2.50 per million input tokens and $10 per million output tokens. That sounds cheap, but long conversations add up fast. Here's a concrete example: if you paste a 5,000-word research document into an AI tool and ask it to summarize, you've just used about 6,700 input tokens before the AI even starts writing. Do that ten times a day and you're burning through serious usage. A practical tip: before pasting a huge document into any AI tool, trim it down. Remove headers, footers, repeated sections, and anything you don't actually need the AI to read. You'll cut your token usage by 20-30% without losing anything important. For a deeper dive, see our guide on [AI content ROI measurement](/blog/ai-content-roi-measurement). **Related**: How do I check how many tokens I'm using? | Why does the same prompt cost different amounts on different days?