Estimate what an API call costs from input and output tokens and the per-million price. Set your own prices or start from a tier, then project it to a month. Runs in your browser.
Almost every LLM API bills per token, with input (your prompt,
history, and retrieved docs) and output (the reply) priced separately, output is usually
several times more expensive. Cost is simply
(input ÷ 1,000,000 × input price) + (output ÷ 1,000,000 × output price). The surprise on most
bills is volume: a fraction of a cent per call becomes real money across thousands of requests a day, which
is what the monthly projection here is for.
The tier presets are round-number bands reflecting common pricing as of June 2026, not a live price sheet for any specific model. The price fields are editable, always confirm the real number with your provider. Nothing you enter leaves your browser.
Generating tokens is more compute-intensive than reading them, so providers price output higher, often 3–5×. That’s why trimming verbose responses saves more than trimming the prompt.
No. They’re illustrative bands to start from, dated and rounded. Specific model prices change often, so the fields are editable and you should verify against your provider.
No. It’s pure arithmetic in your browser, nothing is sent anywhere and it works offline.
Count tokens with the Token Counter, plan the prompt with the Context Window Visualizer, or browse AI Explained.