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

# Models

> OverControl provides a family of proprietary Lume models through one API.

Choose a model based on the capability, speed, modality, and efficiency your application needs. You can change models without changing the rest of your API integration.

<Note>
  Not sure where to start? Use **Lume 3** for most applications. Choose **Lume 3 Max** for the most demanding reasoning and coding tasks, **Lume 2** for lightweight high-volume workloads, or **Lume 3 Omni** when you need multimodal capabilities.
</Note>

***

## Choose a model

<CardGroup cols={2}>
  <Card title="Lume 3" icon="sparkles">
    *Recommended default*

    Strong general-purpose performance with an efficient cost profile. Start here for most text, coding, analysis, and application workloads.
  </Card>

  <Card title="Lume 3 Max" icon="brain">
    *Highest capability*

    The flagship Lume model for advanced reasoning, complex coding, mathematics, physics, and demanding multi-step tasks.
  </Card>

  <Card title="Lume 2" icon="zap">
    *Highest efficiency*

    A cost-efficient text model for lightweight, well-defined, and high-volume workloads.
  </Card>

  <Card title="Lume 3 Omni" icon="shapes">
    *Multimodal*

    A multimodal model with tool use for workflows that need more than text-only processing.
  </Card>
</CardGroup>

***

## Model comparison

| Model       | Model ID    | Modality   | Best for                                                                                                                                                                             |
| ----------- | ----------- | ---------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Lume 3      | lume-3      | Text       | General-purpose applications, coding, analysis, and strong price-performance                                                                                                         |
| Lume 3 Max  | lume-3-max  | Text       | Advanced reasoning, complex coding, mathematics, physics, and difficult multi-step tasks                                                                                             |
| Lume 2      | lume-2      | Text       | Lightweight tasks, high-volume processing, and cost-sensitive workloads Lume 3 Omni	lume-3-omni	Multimodal	Multimodal applications, tool use, agentic workflows, and data extraction |
| Lume 3 Omni | lume-3-omni | Multimodal | Multimodal applications, tool use, agentic workflows, and data extraction                                                                                                            |

***

## Context window

Lume 3-generation models support context windows of up to 400K tokens.

The context window includes the information available to the model during a request, such as instructions, messages, documents, tool context, and generated output.

Large context windows are useful for long documents, codebases, extended conversations, and workflows that need substantial supporting information.

<Warning>
  More context does not automatically produce better results. Send only the information that is relevant to the task whenever possible.
</Warning>

***

## Model IDs

Use the model ID exactly as documented when making an API request.

```text theme={null}
lume-2
lume-3
lume-3-max
lume-3-omni
```

For example:

```json theme={null}
{
  "model": "lume-3"
}
```

To use another model, change the model value. Your API key and Base URL remain the same.

***

## Next steps

<CardGroup cols={3}>
  <Card title="Quickstart" icon="rocket" href="/quickstart">
    Make your first API request with a Lume model.
  </Card>

  <Card title="Pricing" icon="credit-card" href="/pricing">
    Compare pricing across the Lume model family.
  </Card>

  <Card title="Integrations" icon="plug" href="/integrations">
    Connect OverControl to supported tools and OpenAI-compatible clients.
  </Card>
</CardGroup>