# AI models

The selected model answers your chat messages. Its name appears on the left of the header. The app's model selector shows which models are available to your account.

## Select a model

1. Click the model name in the header.
2. Find the model you want in the list.
3. Select it and send your next message.

```{include} ../_includes/ui-models.md
```

This HTML view shows the model catalogue of the documented app version. Your instance may offer a different selection. The model name, context window and capability icons help you choose; the workload indicator can change during operation.

## What should I look for?

| Task | What to consider |
| --- | --- |
| Short questions, rewriting, summarising | A general text model; the preselected model is a useful starting point. |
| Analysing several steps | A model with reasoning capabilities and enough context. |
| Explaining or editing code | A model suited to code; use the [agent](../agents/use-agents.md) to execute it. |
| Analysing images | Use [Computer Vision](use-images.md). Image processing and text responses may use different models. |
| Asking about your own documents | [Activate a project](../arcana/use-arcana-package.md) to include document sources. |

The context window limits how much text a model can process at once. It does not mean that any number of project files will be included in full with every request.

## Adjust responses

Under [Settings → Chat settings](settings.md#chat-settings), you can change **Temperature**, **Top-p**, **Maximum output tokens** and the **System Prompt**. Manage reusable system instructions under [Prompts](prompts.md) in the sidebar.

[Next: Interface overview](interface-overview.md)