# Models

> Choose the right model for your agent extraction tasks.

import AgentWithModelPython from "/snippets/v2/agent/with-model/python.mdx";
import AgentWithModelJS from "/snippets/v2/agent/with-model/js.mdx";
import AgentWithModelCURL from "/snippets/v2/agent/with-model/curl.mdx";

Firecrawl Agent runs on **Spark 2**, the default model for every run. It removes the accuracy-versus-cost decision the Spark 1 models called for: it is cheaper and faster than both of them while matching Spark 1 Pro's accuracy, and it completes a higher share of runs.

## Available Models

| Model | Status | Notes |
|-------|--------|-------|
| `spark-2` | **Default** | Lowest cost, fastest run time, accuracy comparable to Spark 1 Pro |
| `spark-1-pro` | Deprecated | Former default for complex tasks; routes to `spark-2` |
| `spark-1-mini` | Deprecated | Former low-cost Spark 1 option; routes to `spark-2` |

<Note>
**Spark 1 models are deprecated.** The `spark-1-pro` and `spark-1-mini` names remain accepted for backwards compatibility, but every request — with or without the `model` parameter — executes on `spark-2`.
</Note>

## Spark 2

`spark-2` is our newest agent model and powers every agent run.

**Highlights:**
- Lowest cost per run and fastest run time
- Accuracy comparable to the former Spark 1 flagship
- Handles everything from simple lookups to multi-domain research
- The only model with a reasoning budget: pass `effort` (`low`, `medium`, or `high`) to control how hard it thinks

## Spark 1 models (deprecated)

`spark-1-pro` and `spark-1-mini` are deprecated. Their names remain accepted for backwards compatibility, but requests that use them route to `spark-2` — legacy code keeps working, it just runs Spark 2. There is nothing to migrate.

## Specifying a Model

The `model` parameter is optional — every request runs `spark-2`:

<CodeGroup>

<AgentWithModelPython />
<AgentWithModelJS />
<AgentWithModelCURL />

</CodeGroup>

## Pricing by Model

All models use dynamic, credit-based pricing that scales with task complexity:

- **Spark 2**: Uses substantially fewer credits than either Spark 1 model for equivalent tasks
- **Spark 1 models (deprecated)**: Requests route to Spark 2 and are billed accordingly

<Info>
Credit usage varies based on prompt complexity, data processed, and output structure — regardless of model selected.
</Info>

## API Reference

See the [Agent API Reference](/api-reference/endpoint/agent) for complete parameter documentation.

Have questions about which model to use? Email [help@firecrawl.com](mailto:help@firecrawl.com).

> Are you an AI agent that needs a Firecrawl API key? See [firecrawl.dev/agent-onboarding/SKILL.md](https://www.firecrawl.dev/agent-onboarding/SKILL.md) for automated onboarding instructions.