LLM cost per user calculator
Pick a model, enter tokens per request and how often users call it. Prices show the date they were last verified.
AI cost per user per month
$0.43
AI leaves a 97.9% margin at $20/month (before other costs).
- Cost per request
- $0.00142
- Cost per 1,000 requests
- $1.42
- Share of cost from output tokens
- 56.3%
Want this from real data, every day? Connect Stripe and get it automatically →
The formula
Prices are per million tokens. Cached input — repeated context served from a prompt cache — is priced at the cached rate. Multiply cost per request by requests per user per month for cost per user.
Getting realistic inputs
- Log token counts from real API responses rather than guessing.
- Include the system prompt, retrieved documents and conversation history in input tokens.
- Use a heavy-user percentile for requests per month, not the average.
- Count retries, tool calls and agent steps as separate requests.
About the prices
Prices come from the providers' published list prices and show when they were last verified. Providers change prices and add models often — always confirm on the provider's pricing page before making decisions.
Read how to calculate LLM cost per user, then turn the estimate into a price with the AI pricing calculator.
Frequently asked questions
Why does output cost so much more?
Providers price output tokens several times higher than input tokens, so long responses often dominate cost.
How accurate is this?
It is an estimate based on list prices and your assumptions. Real cost depends on actual token counts, retries and any discounts you have.
Can I track real cost per model?
Yes — MarginMeter imports OpenAI and Anthropic cost reports daily by model and project and shows them against your revenue.
Related tools
See your real margin in five minutes.
Connect Stripe and your cost providers with read-only access. Free for two connections — no card required.