Loading page
About this tool
Work out AI cost per active user per month from sessions, message length, conversation history growth and token pricing.
Cost per active user is the AI unit economic that decides whether a feature is priceable: sessions per user × cost per session, where a session costs (prompt tokens × input rate + output tokens × output rate) ÷ 1,000,000. The number most spreadsheets miss is conversation history — a stateless chat API resends the system prompt and the whole transcript on every turn, so prompt tokens grow with turn number and a session's total grows quadratically. For n turns with a system prompt of s tokens, user messages of u and replies of a, total prompt tokens are n·s + u·n(n+1)/2 + a·n(n−1)/2.
Open LLM Cost Per User Calculator on AltFTool — it loads instantly in your browser.
Enter the values you already know.
Fine-tune the options to match your scenario.
Read the result and use it in your planning or reporting.
Models the quadratic token growth of resent transcripts instead of assuming flat per-message cost.
Puts cost per user next to your seat price and reports gross margin and break-even sessions.
Applies a cache hit rate at the discounted cached-input rate, the way prompt caching actually bills.
Multiply cost per session by sessions per user per month. Cost per session is prompt tokens times the input rate plus output tokens times the output rate, both divided by one million. At $3 per million input and $15 per million output, a five-turn session with 8,750 prompt and 2,000 output tokens costs about $0.056.
Because the whole transcript is resent on every turn. Turn 10 pays for turns 1 through 9 again, so prompt tokens grow linearly per turn and the session total grows with the square of the turn count. Doubling conversation length can roughly quadruple the input bill.
For a light assistant on a small model it is often a few cents; for heavy multi-turn use on a frontier model with large context it can exceed several dollars. The spread between the median and the 99th-percentile user is usually the bigger risk to margin than the average itself.
Yes, when the same prefix — system prompt, tool definitions, retrieved documents — repeats across turns, because cache reads are billed far below fresh input tokens. The saving scales with your cache hit rate, so a 70% hit rate on the prompt side removes most of the input cost but none of the output cost.
Add the LLM Cost Per User Calculatorwidget to your blog or website — free, responsive, no signup. Just keep the “Widget by AltFTool” credit link visible.
<iframe src="https://www.altftool.com/embed/widget/llm-cost-per-user-calculator"
title="LLM Cost Per User Calculator — free AltFTool widget"
width="100%" height="640" style="border:0;border-radius:12px;overflow:hidden"
loading="lazy" referrerpolicy="no-referrer-when-downgrade"></iframe>
<p style="font-size:12px;margin:4px 0 0">Widget by <a href="https://www.altftool.com/tools/all/llm-cost-per-user-calculator?utm_source=embed&utm_medium=widget">AltFTool — free online tools</a></p>