Return on Token Spend, RTS Calculator

The Return on Token Spend, RTS calculator is built for AI founders, product managers, CFOs, infrastructure teams, and developers evaluating API-based AI product economics. It combines input token cost, output token cost, platform fees, request volume, and AI-attributed revenue to show gross profit, RTS multiple, RTS percentage, and request-level break-even economics. RTS is especially useful when model API usage is a major driver of product margin and pricing strategy.

Premium planner

Return on Token Spend, RTS Calculator Interactive Tool

Adjust the assumptions below and review the live results locally in your browser.

Results

Enter values to see results

How to use this calculator

  1. Enter the revenue generated from AI usage for the selected monthly, quarterly, or annual period.
  2. Enter input tokens, output tokens, and the price per 1M tokens for each token direction.
  3. Add any other API, orchestration, observability, vector database, or platform costs tied to the AI feature.
  4. Enter the number of AI requests or generations to calculate cost per request, revenue per request, and break-even revenue per request.
  5. Review the RTS multiple, RTS percentage, profit bridge, and token-cost sensitivity chart before changing pricing or usage limits.

The Formula

The mathematical logic powering this calculation is:

Input token cost = input tokens / 1,000,000 × input token price
Output token cost = output tokens / 1,000,000 × output token price
Total token/API cost = input token cost + output token cost + other API/platform costs
Gross profit = revenue generated - total token/API cost
RTS multiple = revenue generated / total token/API cost
RTS percentage = gross profit / total token/API cost × 100
Cost per request = total token/API cost / number of requests

Worked Example

If you input:

  • Revenue Generated from AI Usage: $120,000
  • Input Tokens: 850,000,000
  • Output Tokens: 220,000,000
  • Input Token Price: $2.50
  • Output Token Price: $10.00
  • Other API / Platform Costs: $1,250
  • AI Requests / Generations: 1,800,000

Result:

With the default assumptions, total token/API cost is about $5,575, gross profit is about $114,425, and RTS is roughly 21.52x before other non-AI product costs.

Frequently Asked Questions

What does RTS measure?

RTS measures how much AI-attributed revenue is generated for a given amount of token and API/platform spend. It is a gross AI usage efficiency metric, not a full company profitability metric.

When is RTS most useful?

RTS is most useful for API-based AI products, copilots, agent features, AI search, content generation, support automation, and other products where token usage is a direct variable cost.

Should output tokens be included?

Yes. Output tokens are often more expensive than input tokens, so ignoring them can materially overstate margins and make pricing look safer than it really is.

What should count as revenue generated from AI usage?

Use revenue that can reasonably be attributed to the AI feature, plan, workflow, or usage event. Avoid counting revenue that would have happened without the AI capability.

How should platform fees be handled?

Include other API and platform costs that scale with the AI feature, such as orchestration, vector database usage, evaluation tools, monitoring, routing, or provider minimums.

What does a negative RTS percentage mean?

A negative RTS percentage means token/API cost is higher than AI-attributed revenue for the selected period, so the usage is losing gross profit before considering payroll, support, hosting, and overhead.

Is RTS the same as RAI?

No. RTS evaluates token/API spend efficiency. RAI evaluates owned or leased AI infrastructure economics after operating costs, depreciation, utilization, and payback assumptions.

Similar Calculators