Token Consumption Metrics
Definition
Measures the cost of token-based models (e.g., OpenAI GPT) based on input/output token usage. This KPI helps predict and control costs for LLMs, which charge per token. Facilitates prompt engineering to reduce token consumption without degrading output quality.
Formula, as published
Cost Per Token = Total Cost / Number of Tokens Used
- Data sources:
- API usage reports Dashboards from AI platforms Logs from AI platforms Cloud billing data
- Example:
- If the total cost for inference is $2,500 and the number of tokens processed is 1,000,000, the cost per token is: $2,500/1,000,000 = $0.0025 per token
Calculate it
Example values to start with; replace them with yours.
How to measure it
Cost per token is easier to read per million tokens, which is how providers price. The library’s example, $2,500 for one million tokens, is $2,500 per million: far above today’s list prices, so check the unit before comparing.
Input, output and cached tokens are priced differently, so a blended cost per token moves when the mix does even if no price changed.
Drop in your OpenAI or Claude usage export and see where the AI spend goes.
From the FinOps Foundation
More Reporting & Analytics KPIs
Definition and formula: Token Consumption Metrics, FinOps KPI library, FinOps Foundation, licensed CC BY 4.0, read 16 Oct 2026. Changed: HTML removed, dashes and British spellings changed, formula split from its data sources and examples. The calculator and the notes are Finitizer’s.
Measured once, a KPI is a number. Tracked daily, it is a practice.
Finitizer computes the KPIs your billing data supports every day, by team and product, across AWS and Google Cloud.
See how Finitizer works