Last updated: April 5, 2026 · Core Concepts · by Daniel Ashford

What is Tokens?

QUICK ANSWER

The basic units of text that LLMs process — roughly 3/4 of a word.

Definition

Tokens are the fundamental units of text that language models process. Rather than reading whole words, LLMs break text into smaller pieces called tokens using a process called tokenization. One token is approximately 3-4 characters or roughly 75% of a word in English.

How It Works

Different models use different tokenization methods. GPT models use Byte-Pair Encoding (BPE), while others use SentencePiece or WordPiece. Common words like "the" are usually a single token, while uncommon words may be split into multiple tokens. Token count matters because LLMs have maximum context windows measured in tokens, and API pricing is calculated per token.

Example

"Hello, how are you?" is approximately 6 tokens. A typical page of English text is about 500-700 tokens. GPT-5.3 Codex charges $10 per million input tokens.

Related Terms

Context Window
The maximum amount of text an LLM can process in a single request.
Input Tokens
The tokens in your prompt that the model reads — cheaper than output tokens.
Output Tokens
The tokens the model generates in its response — the most expensive part of API usage.
LLM API Pricing
The cost of using language models, typically measured in dollars per million tokens.

See How Models Compare

Understanding tokens is important when choosing the right AI model. See how 12 models compare on our leaderboard.

View Leaderboard →Our Methodology
← Browse all 47 glossary terms
DA
Daniel Ashford
Founder & Lead Evaluator · 200+ models evaluated