The plain-English definition
A token is the smallest unit of text a language model reads and writes. When you send a prompt, the model does not see words or characters; it sees a sequence of tokens. When it replies, it generates tokens one at a time. Providers bill you per token, so the token is also the unit of cost.
How tokenizers split text
A tokenizer is the algorithm that turns raw text into tokens. Most modern models use a variant
of byte-pair encoding (BPE), which learns a vocabulary of common character sequences. Frequent
words become a single token; rare or long words are broken into smaller pieces. OpenAI models
use a tokenizer called o200k, which this site runs exactly in your browser.
Why tokens are not words
It is tempting to treat tokens and words as the same, but they diverge quickly. The word “unbelievable” is several tokens, while “the” is one. Punctuation, spaces, numbers and emoji all count. Structured text like JSON spends tokens on braces, quotes and repeated keys. This is why a word count is a poor proxy for cost and a token count is the reliable one.
Why token count matters
Token count decides three practical things: whether your prompt fits the model’s context window, how much the request costs, and how fast it responds. Sizing a prompt by words can overflow a context window or blow a budget without warning. Measuring tokens first avoids both.
See your own text
Paste any text below for an exact GPT token count, a token visualization, and a labeled estimate for other models.
Token count uses the exact GPT (o200k) tokenizer. Counted in your browser, never uploaded.
| Model | Tokens | Per request | Monthly | Context |
|---|---|---|---|---|
| Qwen3.8 Max | 0 est. | $0.003 | $3.00 | 0.1% |
| DeepSeek V4 Flashcheapest | 0 est. | $0.00014 | $0.14 | 0.1% |
| Claude Opus 5 | 0 est. | $0.013 | $12.50 | 0.1% |
Token counts vary by tokenizer and model; non-OpenAI counts are estimates. Cost is an estimate and provider pricing can change. See each model's source on the pricing page.