AI Token & Cost Calculator
Paste a prompt, choose a model, estimate output length, and see rough per-run cost. The math is transparent and runs in your browser.
Advanced billing like prompt caching, batch discounts, web search, image generation, and tool calls is not included.
Characters
0
Input tokens
0
One run
$0.0000
Total
$0.0000
How the estimate works
The calculator estimates input tokens from your text length, adds the output tokens you expect, and multiplies both by the selected per-million-token rates. The default text estimate is roughly one token per four characters.
Formula: cost = inputTokens / 1,000,000 * inputRate + outputTokens / 1,000,000 * outputRate. Total cost multiplies that per-request estimate by the number of runs.
Sources
Limitations
This page is for planning, not accounting. It does not perfectly reproduce provider tokenizers, cached-input pricing, batch discounts, reasoning-token billing, retrieval storage, web search fees, tool calls, images, audio, or enterprise discounts.
Last reviewed: 2026-05-14. Pricing presets are manually reviewed and should be checked against official provider pages before budget decisions.
Frequently asked questions
Is this the exact number that my provider will bill?
No. This is an estimate. Real billing can include tokenizer differences, reasoning tokens, tool calls, image/audio costs, cached input discounts, batch pricing, taxes, and provider-specific rounding.
Why does the calculator estimate tokens from characters?
A browser-only static page cannot perfectly match every provider tokenizer. The default estimate uses about four characters per token, which is useful for planning but not a billing source of truth.
Is my prompt uploaded?
No. The token estimate and cost math run locally in your browser. Your pasted text is not sent to a server.