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

# Overview

> What Inference does, and where to go next in this section.

Inference is where you call a model you've already registered in Eval or trained in Train. Every deployed model gets its own OpenAI-compatible endpoint, you authenticate with a key scoped to that model and send it requests directly, from `curl`, your own backend, or any OpenAI-compatible client.

***

## Get a key and call a model

<CardGroup cols={2}>
  <Card title="Inference API Token" icon="key" href="/docs/inference/api-token">
    Issue a key from a model's page, then call its endpoint directly.
  </Card>

  <Card title="BenchGen Router Lite" icon="shuffle" href="/docs/inference/benchgen-router-lite">
    Point that same key at a router that picks the model for you per query.
  </Card>
</CardGroup>