Token

Simple Definition

A token is the basic unit of text that AI language models process. When you send a message to an AI, it doesn’t read word by word — it breaks your text into tokens first.

Tokens are roughly chunks of 3–4 characters. Common words are often one token; rarer words may be split into multiple tokens.

How Tokenization Works

The text “ChatGPT is great” might be split into tokens like:

• “Chat” + “G” + “PT” + ” is” + ” great”

Or possibly: “ChatGPT” + ” is” + ” great”

Tokenization depends on the model’s vocabulary and the specific tokenizer used.

Why Tokens Matter for Users

1. Context window limits — AI models can only process a certain number of tokens at once. Long conversations or documents need to fit within this limit.

2. API pricing — AI APIs charge by the token (both input and output tokens). Understanding token counts helps you estimate costs.

3. Response length — when you ask an AI to “keep it under 200 words,” you’re really controlling token count indirectly.

Rough Token Estimates

Input vs. Output Tokens

• Input tokens — everything you send to the model (your prompt, conversation history, documents)
• Output tokens — the response the model generates

APIs typically charge for both, with output tokens often costing more.

Related Terms

• Context Window — the maximum number of tokens a model can handle at once
• LLM — language models that process text as tokens
• Inference — the process where tokens are generated
• Temperature — the setting that affects token selection during generation