LLM Fundamentals

What is a Token?

A token is the basic unit of text AI processes—approximately 0.75 words or 4 characters. Tokens determine both AI pricing (cost per token) and context limits (how much AI can "remember").

Understanding Tokens

AI doesn't read text the way humans do. Instead of processing individual letters or words, it breaks text into "tokens"—chunks that typically represent parts of words, whole words, or common phrases. This tokenization process is how AI manages and understands language.

Think of tokens as the currency of AI. You "spend" tokens when you write prompts, and the AI "spends" tokens when it responds. Understanding tokens helps you write more efficient prompts and estimate costs when using AI APIs.

How Text Becomes Tokens

Example sentence tokenized:

The beautiful 3 -bed room home features hard wood floors

10 tokens for 9 words (note: "3-bedroom" and "hardwood" split into multiple tokens)

100

words

75

tokens

1

page of text

500

tokens

1

average email

150

tokens

Why Tokens Matter

💰 Pricing

AI APIs charge per token. Both input (your prompt) and output (AI response) cost money.

Example pricing (GPT-4):

  • • Input: $0.03 per 1K tokens
  • • Output: $0.06 per 1K tokens
  • • 1,000 emails ≈ $4.50

📚 Context Limits

Each AI has a maximum context window—the total tokens it can process in one conversation.

Context windows:

  • • GPT-4: 128K tokens (~96K words)
  • • Claude: 200K tokens (~150K words)
  • • Gemini: 2M tokens (~1.5M words)

Tokens for Real Estate Agents

For most agents using ChatGPT Plus, Claude Pro, or similar subscriptions, token limits rarely become a concern for normal use. However, understanding tokens helps you:

Write More Efficient Prompts

Concise prompts leave more room for AI responses. Instead of rambling instructions, be specific and direct.

Understand Context Rot

Long conversations consume tokens. As you approach context limits, AI may "forget" earlier instructions. Start fresh conversations for best results.

Estimate API Costs

Building AI tools for your team? Understanding tokens helps budget API costs. A 500-word listing description costs about $0.03-0.05 to generate.

Frequently Asked Questions

What is a token in AI?
A token is the basic unit of text that AI models process. One token is approximately 0.75 words or 4 characters. When you interact with AI, both your input and the AI's output consume tokens. Tokens determine pricing (you pay per token) and context limits (how much the AI can "remember").
How many tokens is a word?
On average, 1 word equals about 1.3 tokens. Short common words often equal 1 token, while longer or unusual words may be split into multiple tokens. A rough rule: 100 words ≈ 75 tokens, or 1,000 tokens ≈ 750 words.
Do I need to worry about tokens with ChatGPT Plus?
For most normal use, no. ChatGPT Plus, Claude Pro, and similar subscriptions include generous token allowances. The main time tokens matter is during very long conversations (context rot) or when building custom AI applications using APIs.
How can I count tokens in my text?
OpenAI provides a free tokenizer tool at platform.openai.com/tokenizer. You can paste text and see exactly how it's tokenized. For quick estimates, divide word count by 0.75 or multiply character count by 0.25.

Related Concepts

Related Articles

Sources & Further Reading

Learn AI Fundamentals

Understand tokens, context windows, and other AI fundamentals in The Architect workshop—practical knowledge for effective AI use.

View Programs