Find out whether your AI feature makes money on each user. Enter how much a user calls the model and what they pay you, and see the AI cost per user, your gross margin, and how much usage it takes before the feature stops paying for itself.
Indicative model prices in USD, last reviewed —. Keep all amounts in USD to compare cleanly.
AI breaks the SaaS rulebook
Classic software serves one more user for almost nothing. AI doesn't — every call has a real cost, so heavy users can quietly cost more than they pay. Watch the margin on your average user, then the heaviest.
Caps protect the downside
A fair-use limit, credits, or a metered tier stops a power user from turning a profitable plan into a loss. The break-even usage figure tells you where to set the ceiling.
Model choice is the big lever
Switch the model above to see margin swing. A smaller model at a tenth of the price often keeps quality while transforming the unit economics.
The per-user profitability of an AI feature — what one active user costs you in model calls versus what they pay. It's AI's version of cost of goods sold. A $20 user whose AI usage costs $14 leaves a $6 (30%) AI gross margin.
Every AI call has a real marginal cost, so heavy users can cost more than they pay. Knowing cost per user tells you whether price, caps, or model need to change before you scale.
SaaS targets 70–80%; AI features often run lower. Many aim for 50–70%. Below ~40% is fragile, and a negative margin on your average user means the feature loses money on every customer.
Smaller models where quality allows, shorter prompts, capped output, caching, batching, and fair-use limits. Routing easy calls to a cheap model and escalating only hard ones is especially effective.
No. Everything runs in your browser; nothing you enter is sent to a server.
Get an email when we ship the next AI cost or pricing tool.
No spam, no signup needed to use any tool. Unsubscribe any time.