Free calculator · price before you scale

AI pricing calculator: plan price and break-even

Calculate contribution per customer, the minimum price that covers payment and retry-adjusted model cost, and the customers needed to cover fixed cost.

Enter one monthly pricing scenario

All prices, costs, and model rates are in USD.

Model

Monthly price per customer

Monthly fixed costs

Processor fee (%)

Fixed processor fee

AI calls per customer

Extra retry calls (%) Include failed or timed-out calls that are retried.

Uncached input tokens per call

Cache-read tokens per call For OpenAI, subtract these from total prompt tokens before entering uncached input.

5-minute cache-write tokens per call

1-hour cache-write tokens per call

Output tokens per call

Pricing advice needs your economics.

Enter your price, fees, fixed costs, and model usage. The page will not invent a “good” margin for you.

How to read it

Two break-even questions, one answer

The minimum break-even price covers payment fees and retry-adjusted model usage for one customer. Fixed-cost break-even then asks how many positive customer contributions are needed to cover the monthly fixed costs you entered. Neither result includes taxes, refunds, discounts, payroll, or other variable costs.

Calculation method

AI cost per customer = calls per customer × (1 + extra retry percentage / 100) × token cost per attempt. Each separate token bucket is multiplied by its USD rate per million tokens and divided by 1,000,000.

Contribution per customer = plan price − (plan price × processor percentage / 100) − fixed processor fee − AI cost per customer. Break-even customers = monthly fixed costs ÷ contribution per customer, rounded up. Non-positive contribution cannot cover positive fixed costs; the result is unavailable, not a negative customer count.

Worked example · synthetic, USD

Minimum price = (AI cost per customer + fixed payment fee) ÷ (1 − percentage fee ÷ 100). This is the per-customer threshold before fixed monthly costs and other excluded expenses, not a recommended selling price. It assumes one payment per customer per month and a percentage fee below 100%.

GPT-4.1 mini with 1,000 uncached input, 200 cache-read and 100 output tokens costs $0.00058 per attempt at $0.40 / $0.10 / $1.60 per million respectively. 100 calls × 1.05 = 105 attempts per customer, or $0.0609. A $20 plan less 3% fees ($0.60), a $0.30 fixed fee and $0.0609 AI cost leaves $19.0391 per customer. Covering $100 monthly fixed costs needs 6 paying customers.

These are entered assumptions, not accounting profit or a forecast. The fixed payment fee is modeled once per customer per month. Refunds, taxes, discounts and other variable expenses are not modeled; subtract them separately before making a pricing decision. Stress-test the whole period when actual usage differs.

In that scenario, the per-customer threshold is ($0.0609 + $0.30) ÷ 0.97 ≈ $0.372062. If charging whole US cents, $0.37 still loses money on those costs; $0.38 is the first cent-denominated price above the threshold. Neither price recovers the $100 fixed cost by itself. Load this synthetic example to inspect or change its assumptions.

Price basis and limits

These are dated catalog rates, not a live price feed. OpenAI and Anthropic entries were recorded on 2026-07-28; BytePlus Seed 2.0 Lite on 2026-07-13. Check the selected model's source and verification date shown with the result before budgeting. Re-running this page does not update those dates.

Only supported token buckets are priced. Models without a separate cache-write rate reject cache-write input. Seed 2.0 Lite rejects more than 128,000 input-side tokens per request because the higher pricing tier is not modeled. No FX conversion is applied.

Provider references: OpenAI pricing, Anthropic pricing, and how Grow or Die versions prices. This tool is an estimate, not a reconciled provider bill.