What it does
The AI Pricing Calculator models how an AI product’s pricing actually performs — on margin, not just on the page. Set the economics that matter — your cost per unit of usage, how heavily customers use the product, the included allowance, the top-up price, and a good/better/best tier ladder — and it returns the blended gross margin across your tier mix, each tier’s economics, the customer’s bill (and the bill shock when usage spikes), CAC payback, LTV, and where your sweet-spot price sits.
Start from scratch, or load a real company’s published pricing straight from the AI Credit Index — Cursor, Lovable, Midjourney, and 50+ more — and stress-test its structure under your own cost and usage assumptions. Prices load from the companies’ own public pages; the costs and usage are yours, so this models how their pricing performs under your numbers — never a claim about their actual margins.
What you can pressure-test
The main Stress Test is a full diagnostic, not a single number. Pick a strategy mode — Market Capture, Balanced Growth, Monetization, or your own custom target bands — and every output is graded against it:
- Blended margin & verdict — gross margin and revenue blended across your good/better/best mix, with a one-line read on whether the pricing holds up under your strategy.
- Per-tier economics — revenue, margin, and credit pressure for each tier, plus how many users you’d expect to exceed their included allowance.
- The customer’s bill — what a light user and a heavy user actually pay, and the bill shock when usage spikes.
- Cost breakdown — AI/inference cost, free-tier subsidy, and infra/support, so you can see which users define your margin.
- Margin levers & a solver — a sensitivity (tornado) view of which inputs move margin most, plus a target-margin solver that back-solves the minimum price, maximum included credits, or cost ceiling you need to clear your floor.
- Price what-if & sweet spot — move the price up or down and watch margin, adoption, and revenue respond, with a scan for the revenue-maximizing price.
- Growth economics — CAC payback, LTV, and how the unit economics compound as you grow.
Design and compare structures
- Structures lays the good/better/best ladder out as a billing geometry — included allowances, top-up rates, and tier spacing — so you can design the ladder itself, not just test a single tier.
- Compare puts two or three saved scenarios side by side — outputs, verdicts, and exactly which inputs differ — so you can settle which configuration wins with the numbers in front of you.
Save, share, and export
Save any configuration as a named scenario, share it as a link that reopens the exact state, and export a clean PDF of the full diagnosis for a deck or a pricing review. Everything runs locally in your browser — your cost and usage numbers never leave it.
Who it’s for
Founders pricing their first AI feature. Operators bolting usage-based AI onto an existing seat-based model. Anyone who needs to know whether a credit allowance, a top-up rate, or a three-tier ladder still holds up once real usage and real cost hit it.
How to use it
Set your cost per unit and a usage profile — or load a company to prefill its published prices. Tune the included allowance, the top-up price, and up to three tiers, pick a strategy mode, and read the verdict. Then push on it: move the price, solve for your margin floor, compare a second structure side by side, and export the result.
Why this exists
Pricing AI is the hardest pricing problem in software right now. Usage is genuinely variable, the unit costs aren’t fully understood, and the headline price hides the few heavy users who define your margin. Benchmarks tell you what everyone else charges; this calculator answers the other half — what you should charge, and what it costs you to deliver.