Skip to content

Token

1. What Is It?

One-sentence definition

A Token is the smallest text-processing unit used by an AI model, and it is also the basic unit used for pricing. It is not exactly the same as a word or a character. It is closer to a semantic fragment.

2. How It Works

The model cannot read plain text directly. It first converts text into token IDs through tokenization.

3. Tokens and Pricing

  • Chinese often consumes more tokens
  • English is often more token-efficient
  • both input and output are usually billed

4. Why Long Conversations Become Expensive

Because multi-turn systems often resend earlier history along with the new message, token consumption grows much faster than users expect.

5. Intuitive Analogies

  • tokens are like puzzle pieces
  • tokens are like a taxi meter
  • the context window is like a fuel tank

6. Practical Ways to Save Tokens

MethodExample
Ask more directlyRemove unnecessary filler
Limit answer length"Answer in under 100 words"
Start a new conversation when neededAvoid endlessly growing history
Use summariesCompress earlier turns
Choose the right modelCheap model for simple tasks

7. What You Need to Remember

  • tokens are the unit of model text processing and billing
  • both cost and context length depend on tokens
  • a large amount of AI engineering is really about using limited tokens more efficiently

From Zero, To Next