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

# Models and modes

# Models and Modes Reference

Stable selection guidance for Exa search types and compatibility-layer model routing.

## Canonical Docs Links

* Base docs URL: `https://exa.ai/docs`
* Search reference: `/reference/search`
* OpenAI SDK compatibility: `/reference/openai-sdk`

## Search Types

Exa's primary search modes for new work:

| Type             | Use When                                            | Tradeoff                            |
| ---------------- | --------------------------------------------------- | ----------------------------------- |
| `auto`           | You want the default                                | Best general-purpose starting point |
| `fast`           | Latency matters more than maximum synthesis quality | Lower latency                       |
| `instant`        | Real-time UI and assistant flows                    | Lowest latency                      |
| `deep-lite`      | You want lightweight synthesis                      | More reasoning than default modes   |
| `deep`           | You need deeper synthesized retrieval               | Higher latency                      |
| `deep-reasoning` | You want the strongest research-style reasoning     | Highest latency                     |

## How Output Controls Affect Behavior

* `outputSchema` adds synthesis work and therefore adds latency across search types
* `systemPrompt` guides synthesis behavior but does not replace schema control
* `maxAgeHours: 0` forces fresh crawling and can increase latency substantially

Do not choose a deep variant only because you want structured output. Choose it when the retrieval task itself needs more reasoning or multi-step synthesis.

If you are optimizing for responsiveness, start with:

* `type: "fast"` for coding or agent workflows, `type: "instant"` for real-time UX, or `type: "auto"` for general retrieval
* no `outputSchema`
* default freshness behavior

Then add deeper reasoning or stronger freshness only when the use case requires it.

## Answer and Compatibility Models

In OpenAI-compatible flows, current model routing is:

| Model | Typical Use                                        |
| ----- | -------------------------------------------------- |
| `exa` | Answer-style compatibility via `/chat/completions` |

These are compatibility-layer names. Native Exa endpoint choice remains the more important design decision.

Use `deep-reasoning` when:

* the task needs multi-step synthesized retrieval
* you want a current Exa-first path instead of the legacy research surface
* you need a replacement for legacy `/research/v1` structured output flows

## Practical Defaults

* General retrieval: `type: "auto"`
* Fast coding or agent path: `type: "fast"` with `highlights`
* Real-time UX: `type: "instant"` and minimal synthesis