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

# Choose the AI model

> Pick the AI model that writes your answers, and split your message quota between the models.

Your chatbot can answer with one of two AI models:

| Model            | Use it when                                                                          | Message cost             |
| ---------------- | ------------------------------------------------------------------------------------ | ------------------------ |
| **GPT-4.1 Mini** | Most chatbots. Fast answers from clear content.                                      | 1 message                |
| **GPT-4.1**      | Answers must follow long or detailed instructions closely, or questions are complex. | 10 GPT-4.1 Mini messages |

One GPT-4.1 message uses as much of your quota as 10 GPT-4.1 Mini messages. New chatbots use GPT-4.1. For most chatbots, switch to GPT-4.1 Mini. Use GPT-4.1 only if answers do not follow your instructions well enough.

For the number of messages on each plan, see [Plans and limits](/docs/reference/plans-and-limits).

## Before you start

* **Role:** Manager or higher to change the model. See [Team members](/docs/guides/account/team-members).
* The quota split belongs to the account that owns the chatbot. Sign in as the owner to change it on the **Usage** page.
* **Plan:** any plan.

## Change the model

<Steps>
  <Step title="Open General settings">
    Go to **Settings** > **General**.
  </Step>

  <Step title="Select the model">
    In **GPT Model**, select **GPT-4.1 Mini** or **GPT-4.1**.
  </Step>

  <Step title="Save">
    Select **Save Changes**.
  </Step>
</Steps>

The model applies to this chatbot only. Each chatbot has its own model setting.

## Split your message quota between the models

Your monthly messages are shared between the two models. On the **Usage** page, you choose how much of your quota goes to each model.

<Steps>
  <Step title="Open Usage">
    In the top navigation, select **Usage**.
  </Step>

  <Step title="Move the slider">
    Move the slider to split your quota. The page shows the result as **GPT-4.1 messages**, **GPT-4.1-mini messages**, and **Total messages**.
  </Step>

  <Step title="Save">
    Select **Save Changes**.
  </Step>
</Steps>

Things to know about the split:

* Moving quota to GPT-4.1 gives you fewer total messages, because each GPT-4.1 message costs 10 GPT-4.1 Mini messages.
* You cannot set a share lower than what you already used this month.
* If you have used more than your whole monthly quota, you cannot save a new split. The page says "You've exceeded your combined monthly quota, so there's no valid split to save." Wait for the monthly reset, or [add messages or upgrade](/docs/guides/account/billing).

## When one model's share runs out

When the share for your chatbot's model is used up, the chatbot answers with the other model until the monthly reset. This applies on your website, in your integrations, and through the API, the CLI, and MCP. An open chat on your website keeps its model until the visitor loads the page again. For example, if your GPT-4.1 share is used up, the chatbot answers with GPT-4.1 Mini. When both shares are used up, [add messages or upgrade your plan](/docs/guides/account/billing).

## Check that it works

Open the chat button in the corner of your dashboard and start a new conversation. Ask a question.

Then go to **Chat History**, open **Filter**, and select your model under **GPT Model**. Your new conversation shows in the list.