> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.sparkian.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.sparkian.com/_mcp/server.

# Multi-model chat

Instead of picking a single model, you can send the same message to several models at once and see each response side by side.

## Turning it on

Open the model picker and toggle **Multi-model chat**. A comparison set of models is preselected by default whatever your [workspace's comparison models](/workspaces) are set to and you can add or remove models from the set before sending.

![Multi-model chat toggle](https://cdn.sparkian.com/docs/multi-model-toggle.png)

## Comparing responses

Each model gets its own column with independent controls:

![Side-by-side model responses](https://cdn.sparkian.com/docs/multi-model-responses.png)

* **Copy** or **Download** a response individually
* **Retry this model** to regenerate just that response
* **Delete response** of specific model
* **Hide response** to collapse a column without removing it
* **Expand to full width** to focus on one response
* **Compare with another model** to add a model to the set mid-conversation
* **Make this the primary response** to designate which response the conversation continues from if you switch back to single-model chat

Each model in the comparison charges its own Sparks. These per-column actions are the multi-model version of the standard [chat response actions](/chat-response-actions).

## Related

* [Chat response actions](/chat-response-actions)
* [Chat branching](/chat-branching)
* [Workspaces](/workspaces)