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

# outerbounds mc pull

export const GCell = ({children, className}) => <div className={`grid-table-cell ${className || ""}`} role="cell">
    {children}
  </div>;

export const GTH = ({children, className}) => <div className={`grid-table-th ${className || ""}`} role="columnheader">
    {children}
  </div>;

export const GRow = ({children}) => <div className="grid-table-row" role="row">{children}</div>;

export const GBody = ({children}) => <div className="grid-table-body" role="rowgroup">{children}</div>;

export const GHead = ({children}) => <div className="grid-table-head" role="rowgroup">{children}</div>;

export const GTable = ({children, className, cols}) => <div className={`grid-table not-prose overflow-hidden rounded-2xl ${className || ""}`} style={{
  "--grid-table-cols": cols
}} role="table">
    {children}
  </div>;

Downloads a model from the [model catalog](/docs/platform/concepts/what-is-the-model-catalog) and prints the absolute path to the downloaded files. Download a model to run it locally with your tool of choice, or to load it into a service you deploy on the platform.

For GGUF models, specify exactly one of `--file` or `--quant`. These options do not apply to safetensors models.

Model downloads are governed by your perimeter's model [policies](/docs/platform/concepts/what-is-a-policy). If a policy denies access to a model, the command fails with an access denied error. Contact your administrator to request access.

## Usage

```bash theme={null}
outerbounds mc pull [OPTIONS] --model <MODEL_NAME> --format [safetensor|gguf] --output-dir <OUTPUT_DIR>
```

## Options

<GTable cols="30% 70%">
  <GHead>
    <GRow>
      <GTH>Option</GTH>
      <GTH>Description</GTH>
    </GRow>
  </GHead>

  <GBody>
    <GRow>
      <GCell>`-m`, `--model`</GCell>
      <GCell><Badge size="sm" color="blue" shape="pill">Required</Badge> Catalog model name (Example: Qwen2.5-7B-Instruct).</GCell>
    </GRow>

    <GRow>
      <GCell>`--format`</GCell>
      <GCell><Badge size="sm" color="blue" shape="pill">Required</Badge> Model format to download.</GCell>
    </GRow>

    <GRow>
      <GCell>`--file`</GCell>
      <GCell>Exact GGUF filename to download.</GCell>
    </GRow>

    <GRow>
      <GCell>`--quant`</GCell>
      <GCell>GGUF quantization (Example: q4\_k\_m, q8\_0)</GCell>
    </GRow>

    <GRow>
      <GCell>`-o`, `--output-dir`</GCell>
      <GCell><Badge size="sm" color="blue" shape="pill">Required</Badge> Directory where the model is stored.</GCell>
    </GRow>
  </GBody>
</GTable>

## Examples

**Download a GGUF model**

```bash theme={null}
outerbounds mc pull --model "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B" --format gguf --quant "q4_k_m" --output-dir .
```

**Download a safetensors model**

```bash theme={null}
outerbounds mc pull --model "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B" --format safetensor --output-dir .
```

**Capture the path in a script**

```bash theme={null}
MODEL_PATH=$(outerbounds mc pull --model "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B" --format gguf --quant "q4_k_m" --output-dir /tmp/model-downloads)
llama-server -m "$MODEL_PATH"
```