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

# DeepSeek

> Use your DeepSeek API key with Rumus to run DeepSeek V3.2 and the V3.2 Reasoner.

Bring your own DeepSeek API key to use the V3.2 family — including the Reasoner model with extended thinking. Requests go directly from your machine to DeepSeek.

## Before you start

You need:

* A DeepSeek account at [platform.deepseek.com](https://platform.deepseek.com/).
* An API key from the **API Keys** page on your DeepSeek dashboard.
* Sufficient credit on your DeepSeek account.

## Add DeepSeek in Rumus

<Steps>
  <Step title="Open the model settings">
    Go to **Settings → AI → Models** and click **Add Model**.
  </Step>

  <Step title="Pick the provider">
    Set **Provider** to **DeepSeek**.
  </Step>

  <Step title="Paste your API key">
    Paste the key into **API Key**. It's stored encrypted in your local [vault](/docs/account/vault).
  </Step>

  <Step title="(Optional) Custom base URL">
    Leave **Base URL** blank to use the default `https://api.deepseek.com`. Override only for proxies or alternative regional endpoints.
  </Step>

  <Step title="Pick a model">
    Choose from the list, or toggle **Enter custom ID** to type a model ID manually.
  </Step>

  <Step title="Capabilities">
    Capability flags are pre-set for the built-in models. For a custom model ID, set them to match what DeepSeek's docs say the model supports.
  </Step>

  <Step title="Save">
    The model appears in the picker under **Custom Models**.
  </Step>
</Steps>

## Recommended models

| Model                      | ID                  | Good for                                              |
| -------------------------- | ------------------- | ----------------------------------------------------- |
| **DeepSeek V3.2**          | `deepseek-chat`     | Default — fast, strong at code, supports tool calling |
| **DeepSeek V3.2 Reasoner** | `deepseek-reasoner` | Longer thinking budget for harder reasoning and math  |

Both models offer a 128K context window. The Reasoner exposes thinking tokens; tool calling is only supported on `deepseek-chat`.

For the latest on the DeepSeek lineup, see the [DeepSeek API docs](https://api-docs.deepseek.com/).

## Tips

* **Reasoner vs Chat.** Pick the Reasoner only when extra thinking helps — it spends more output tokens. For agentic work that needs tool use, stay on `deepseek-chat`.
* **Pricing is among the lowest of any major provider.** If cost matters, DeepSeek is a strong daily driver.

## Troubleshooting

<AccordionGroup>
  <Accordion title="401 Unauthorized">
    Recreate the API key in the DeepSeek dashboard. Make sure it hasn't been revoked.
  </Accordion>

  <Accordion title="Model not in the list">
    Toggle **Enter custom ID** and paste the exact model ID (e.g. `deepseek-reasoner`).
  </Accordion>

  <Accordion title="Tool calls don't work on the Reasoner">
    The Reasoner doesn't support tool calling — switch to `deepseek-chat` for agentic tasks.
  </Accordion>
</AccordionGroup>

<Note>
  Hit a snag we didn't cover? Ask in the [Rumus community](https://www.rumus.ai/community).
</Note>

## Next steps

<CardGroup cols={2}>
  <Card title="Other providers" icon="grid" href="/docs/models/overview">
    Anthropic, OpenAI, Google, Z.AI, Kimi, Ollama, OpenAI-compatible.
  </Card>

  <Card title="Built-in models" icon="cube" href="/docs/models/built-in">
    Use Rumus's bundled DeepSeek models without managing a key yourself.
  </Card>
</CardGroup>