How Many Pages Is 1 Million AI Tokens?
One million tokens is about 1,500 single-spaced pages. At roughly 500 words per page, 750,000 words fill a stack of paper 15 cm (6 in) tall.
Why this matters
Understanding token-to-word conversions is essential for developers budgeting API calls and for businesses estimating costs before committing to an AI provider. When every API call is metered by the token, knowing exactly how far your budget stretches can mean the difference between a viable product and an unexpectedly expensive one.
1 block = 1 page
The answer
≈ 1,500 pages
single-spaced pages
The math
Basis assumption
500 words per page
Calculation
At about 500 words per single-spaced page, 750,000 words fills roughly 1,500 pages — a stack of paper about 15 cm (6 in) tall.
Result
1,000,000 tokens = 1,500 single-spaced pages
Pages by format
"A page" means different things. Plain text is cheap; a PDF read with its layout and images costs about twice as much per page.
| Page type | Words/page | Tokens/page | Pages in 1M tokens |
|---|---|---|---|
| Single-spaced text | 500 | ≈ 667 | ≈ 1,499 |
| Double-spaced manuscript | 250 | ≈ 333 | ≈ 3,003 |
| PDF, read with layout and images | ~400 | ≈ 1,500 | ≈ 667 |
| PDF, text extracted only | ~400 | ≈ 500 | ≈ 2,000 |
| Code file (~30 lines) | — | ≈ 300 | ≈ 3,333 |
| Slide (~50 words) | ~50 | ≈ 67 | ≈ 14,925 |
Cost per page on current models
Sending one single-spaced page (≈ 667 tokens) as input.
| Model | Per page | 1,500 pages (1M tokens) |
|---|---|---|
| GPT-6 Luna | <$0.0001 | $0.10 |
| Gemini 3.5 Flash-Lite | $0.0002 | $0.30 |
| Gemini 3.8 Flash | $0.0005 | $0.75 |
| Claude Haiku 4.5 | $0.0007 | $1 |
| GPT-6 Sol | $0.0013 | $2 |
| Gemini 3.1 Pro | $0.0013 | $2 |
| Claude Sonnet 5 | $0.0013 | $2 |
| Claude Opus 5.5 | $0.0027 | $4 |
| Claude Opus 5 | $0.0033 | $5 |
| GPT-6 Astra | $0.0067 | $10 |
| Claude Fable 5.1 | $0.0067 | $10 |
Quick reference
Other ways to measure one million tokens in the same category:
Real-world references
To put this in perspective, here are some well-known examples for comparison.