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

# Deep Intelligence (Research) mode

> Deep Intelligence is Chief's most powerful response mode.

<Warning>
  ⚠️ **Important:** Deep Intelligence is a **paid-plan feature**. It isn't available on the free plan. If you're on a free plan, use [Automatic, Fast, or Expert](/help/answer-depth-chiefs-agents/response-modes-automatic-fast-expert) — or [upgrade to Pro](/help/plans-credits-billing/plans-overview-explore-pro-power-user-enterprise) to unlock Deep Intelligence.
</Warning>

## The deepest thinking Chief can do

<Frame>
  <img src="https://mintcdn.com/chief-2d405d1a/BBGD1uH7sfT84ugx/images/help/gs-intelligence-levels.png?fit=max&auto=format&n=BBGD1uH7sfT84ugx&q=85&s=e0b3876a72fee1b8c84a2cba58ce83db" alt="The response modes, with Deep Intelligence as the deepest option" width="1300" height="601" data-path="images/help/gs-intelligence-levels.png" />
</Frame>

**Deep Intelligence** is Chief's most powerful response mode. Instead of answering with a single thread of reasoning, it deploys [a fleet of long-running agents](/help/answer-depth-chiefs-agents/meet-chiefs-agents-explorer-librarian-researcher-planner) across multiple leading models that collaborate to get your job done. Chief's orchestrator breaks your request into pieces, runs specialized agents in parallel, and synthesizes their findings into one cited answer.

Use Deep Intelligence when a question is big enough that a single pass won't do it justice — a competitive teardown, a strategy built from internal and external data, a [deep research report](/help/creating-exporting/generating-reports-documents), or anything you'd otherwise assign to a team.

## What makes it different

* **Multiple agents, working in parallel** — Chief spins up as many agents as the task needs, each with its own large context window, rather than reasoning in one pass.
* **Multiple models** — agents draw on the world's most capable models so different parts of the job get the right kind of thinking.
* **Built for longer jobs** — it's designed for deep, multi-step work, so it takes longer than Fast or Expert and is worth reserving for hard questions.

## How to use Deep Intelligence

<Frame>
  <img src="https://mintcdn.com/chief-2d405d1a/ZjT7ckngTj8rjHF7/images/help/answerdepth-deep-intelligence.png?fit=max&auto=format&n=ZjT7ckngTj8rjHF7&q=85&s=fd350ddeda326b69376496cc3c529ccd" alt="The Deep Intelligence option in the mode selector, noting it uses about 5x the credits" width="1200" height="350" data-path="images/help/answerdepth-deep-intelligence.png" />
</Frame>

1. In the [prompt bar](/help/the-prompt-bar-scope/the-prompt-bar-a-tour-of-your-chat-controls), click the **mode selector**.
2. Select **Deep Intelligence**.
3. Type your request and send it.

You can also trigger Deep Intelligence automatically: when you [**mention an agent**](/help/answer-depth-chiefs-agents/calling-on-a-specific-agent-in-your-prompt) (such as **@explorer** or **@librarian**) in your prompt, Chief's router switches that prompt to Deep Intelligence mode for you.

<Warning>
  ⚠️ **Important:** Deep Intelligence uses [roughly 5x the credits](/help/answer-depth-chiefs-agents/how-modes-affect-credit-usage) of a standard response because it runs many agents at once. Reserve it for jobs that genuinely need the depth.
</Warning>

## FAQ

<AccordionGroup>
  <Accordion title="I selected Deep Intelligence but got a message that it requires a paid subscription. Why?">
    Deep Intelligence isn't available on the free plan. To stay on free, switch your prompt to Automatic intelligence; to keep using Deep Intelligence, upgrade to Pro.
  </Accordion>

  <Accordion title="I added an agent to my prompt and the mode jumped to Deep Intelligence. Is that expected?">
    Yes. Mentioning an agent requires Deep Intelligence, so Chief's router switches that prompt to it automatically. If you'd rather not run Deep Intelligence, remove the agent mention and the prompt returns to a standard mode.
  </Accordion>

  <Accordion title="Why is Deep Intelligence slower than Expert?">
    It's running a fleet of agents in parallel and synthesizing their results, rather than answering in a single pass. That extra coordination takes more time — and that's what produces the deeper, more complete answer.
  </Accordion>
</AccordionGroup>