● Runs locally in your browser

Tokens to Words Calculator

Convert a token budget into rough English word count, document pages, and CJK character capacity for context planning.

Private: client-side only.Free: no account or API key.Pricing data verified 2026-08-23
Typical English words75,000
Practical English range65,000–85,000
Approx. 500-word pages150
Approx. CJK characters95,000
Rule of thumb

Tokens are not words.

A token can be a whole word, part of a word, punctuation, whitespace, or a character. The conversion changes with language, formatting, and tokenizer.

For English planning, about 0.75 words per token is a useful midpoint—not a billing-grade conversion.

1,000 tokens≈ 750 typical English words.
100,000 tokens≈ 75,000 typical English words, or roughly 150 pages at 500 words each.
Need the reverse?Paste the actual document into the AI Token Counter for a text-based estimate.
Words to tokens, by content typeProse, code, JSON and HTML convert at different rates — plus what 1,000 words costs on each model: How Many Tokens in 1,000 Words?.
What 100K tokens is~75,000 words, 150 pages, and the chunk size that stays under OpenAI's 272K surcharge: How Many Words Is 100K Tokens?.
Context windows explainedInput and output share one budget — window sizes, output caps and what it costs to fill one: What Is a Context Window?.
What a token isThe unit behind every estimate — ~4 characters, 0.75 words: What Is an AI Token?
Sending scraped HTML?Measure what cleaning it saves: HTML Token Savings.
1M tokens in words750,000 words, 1,500 pages — and why only 5 of 8 models accept it in one request: How Many Words Is 1M Tokens?
Trim the promptFind repeated instructions, markup and filler: Prompt Weight Analyzer
FAQ

Tokens-to-words questions

How many words is 1,000 tokens?

About 750 English words is a useful rough estimate, with substantial variation by text and tokenizer.

How many pages is 100K tokens?

Using 75,000 English words and 500 words per page gives roughly 150 pages.

Does code follow the same conversion?

No. Code, JSON, tables, and heavily punctuated text can tokenize very differently from normal prose.