Free · Private · No signup
Turn tokens into an operating budget. Compare input, cache, output, volume, and context usage across leading models—or enter your own rates.
Pricing checked
August 9, 2026
Against official provider documentation. Prices change; verify before procurement.
The formula
The calculator keeps the arithmetic inspectable and calls out costs that do not fit a basic token formula.
((input × input rate) + (cached × cache rate) + (output × output rate)) ÷ 1,000,000 × requestscharacters per estimated token
separate billing lanes
text sent to our server
Provider tokenizers differ. System instructions, conversation history, schemas, retrieved documents, reasoning, long-context tiers, caching storage, batch discounts, tools, media, and taxes can change the final invoice.
Official pricing sources
OpenAI API pricing
Last verified 2026-08-09
View official source ↗Anthropic API pricing
Last verified 2026-08-09
View official source ↗Google API pricing
Last verified 2026-08-09
View official source ↗Claude Sonnet 5: Introductory price through August 31, 2026; verify before budgeting beyond that date.
Questions people ask
For typical English prose, one token is often about three-quarters of a word, but code, punctuation, languages, and each provider's tokenizer change the result. This calculator uses characters only for a clearly labeled estimate; enter API-reported usage for exact billing.
Multiply uncached input, cached input, and output token counts by their respective per-million-token prices, add the three costs, then multiply by request volume.
They can. Providers may bill hidden or visible reasoning tokens as output tokens. Use the total billed output reported by your provider rather than only the words visible to the user.
No. Those can use separate token rates or per-call charges. The calculator focuses on standard text token costs and explicitly excludes add-on charges.