Usage parameters
Per user economics
Messages/month-
Input tokens/month-
Output tokens/month-
AI cost / user / month-
AI as % of your price-
Total monthly at 1,000 MAU
Total AI cost-
Total revenue-
Gross profit-
AI margin-
How much does your AI feature actually cost per user per month? Enter your usage stats below.
This calculator turns per-user LLM usage into a monthly bill. It multiplies your messages per user per day by average input and output tokens, scales that to a 30-day month, and applies the selected model's per-million-token input and output rates to produce a cost per user per month. It then multiplies by monthly active users for total API spend, and compares the result against your target subscription price to show gross margin per user. This matters because output tokens usually cost several times more than input tokens, and because API cost scales linearly with usage while subscription revenue does not — a small group of heavy users can turn a profitable price point into a loss. Running the numbers before launch shows whether your pricing, model choice, or context size needs to change.
Most providers charge more per output token than per input token, often three to five times more. If your app generates long responses, output tokens can dominate the bill even when each request sends a large prompt or conversation history.
No — it uses the average input tokens you enter for every message. If your app resends the full conversation on each turn, raise the average input token figure to reflect the typical prompt size, not just the newest user message.
Common levers are selecting a cheaper or smaller model for routine requests, trimming or summarizing conversation history, capping maximum output length, and using prompt caching for stable system prompts. Re-run the calculator with the adjusted numbers to see the effect on margin.