About TokenBill

What this is

TokenBill is an independent tool that answers one question: what does a language model actually cost you per month for a specific job? Token prices alone are misleading — the same customer-support reply can cost many times more on one model than another, and prompt-caching hit rates change the math completely. TokenBill combines your workload (requests, input/output tokens, cache hit rate) with current per-token prices to show the real monthly number, side by side. New to cost estimation? Start with the LLM API cost guide.

How the calculation works

For each model, the monthly estimate is:

(input tokens × input price × (1 − cached share × cache discount)) + (output tokens × output price)

All computation happens in your browser. Your inputs are never sent to any server — there is no backend, no account, and no tracking of what you type.

Where pricing data comes from

Model prices are collected from providers' public pricing pages and reputable public sources, and refreshed regularly (see the "Prices updated" date at the top of every page). Prices change — providers update rates, introduce cached-input discounts, and rename models. TokenBill is a planning aid, not a billing authority: always verify current rates with the provider before making purchasing decisions.

TokenBill is not affiliated with, sponsored, or endorsed by OpenAI, Anthropic, Google, xAI, DeepSeek, Alibaba, MiniMax, Moonshot, Zhipu, or any other model provider. Model names and trademarks belong to their respective owners.

Corrections

Spotted an outdated price or a wrong calculation? Corrections are welcome and help everyone. Use the contact details in the site footer to reach us.