> ## Documentation Index
> Fetch the complete documentation index at: https://docs.opsbeacon.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Token Usage & Unit Economics

> Understand the cost per unit of useful work your AI workloads are producing.

## Unit economics for AI

Raw token spend is hard to act on. Unit economics — cost per API call, cost per successful completion, cost per user request — give you a number you can improve.

OpsBeacon calculates unit economics for each AI project where usage data is available.

## How to use it

```
@Beacon what's the cost per API call for the summarization project?
@Beacon show me unit economics for all AI projects
@Beacon which project has the worst cost per completion?
```

## Key metrics

* **Cost per 1K tokens** — blended input + output token cost
* **Cost per API call** — average spend per request
* **Input/output ratio** — prompt tokens vs completion tokens (high input ratios often indicate prompt inefficiency)
* **Average completion length** — longer completions cost more; are they justified?


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.