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

# Agent Model Routing

> Choose how agents select AI models. Smart routing vs specific model selection.

Configure which AI model powers your agent. Choose between platform-managed Smart Routing or specific models.

## Quick Comparison

| Strategy     | Best For          | Flexibility          |
| ------------ | ----------------- | -------------------- |
| **Smart**    | Auto-optimization | High (auto-upgrades) |
| **Specific** | Consistency       | Low (fixed model)    |

## Smart Routing

Platform automatically selects the best model per-request.

**Configuration:**

1. **Agents** > Select Agent > **Settings** > **Model Configuration**
2. Choose **Smart Routing**
3. Optionally set:
   * **Tier** (Budget/Standard/Advanced)
   * **Provider** (OpenAI/Anthropic)
4. Click **Save**

Platform will:

* Analyze request complexity
* Select optimal model
* Auto-upgrade to new models
* Stay within your tier preference

## Specific Model

You choose exact model(s). Optionally add budget model for cost optimization.

**Configuration:**

1. **Agents** > Select Agent > **Settings** > **Model Configuration**
2. Choose **Specific Model**
3. Enter model name (e.g., "gpt-5.1")
4. Optional: Add budget model (e.g., "gpt-4o-mini")
5. Click **Save**

## Available Models

| Tier         | OpenAI      | Anthropic     |
| ------------ | ----------- | ------------- |
| **Budget**   | gpt-4o-mini | claude-haiku  |
| **Standard** | gpt-4o      | claude-sonnet |
| **Advanced** | gpt-5.1     | claude-opus   |

See Console for the complete and current model catalog, or the [Model Routing Platform Guide](/docs/run/platform/model-routing) for full details.

## Model Settings

Fine-tune model behavior per-agent:

```json theme={null}
{
  "temperature": 0.7,
  "top_p": 0.95,
  "max_tokens": 2048
}
```

| Parameter   | Effect                                   |
| ----------- | ---------------------------------------- |
| temperature | Randomness (0=deterministic, 1=creative) |
| top\_p      | Diversity (0.95 is good default)         |
| max\_tokens | Maximum response length                  |

## Budget Models

Use cheaper model for simple requests, premium for complex:

* Primary: gpt-4o (\$expensive)
* Budget: gpt-4o-mini (\$cheap)

Platform automatically routes simple requests to budget model.

Savings: 50-80% on routine queries.

## Fallback

Enable automatic fallback if chosen model unavailable:

* Checked: Use alternative model from same provider
* Unchecked: Fail if model unavailable (not recommended)

See [Model Routing Platform Guide](/docs/run/platform/model-routing) for detailed documentation.

## Next Steps

* [Creating an Agent](/docs/run/agents/creating-an-agent)
* [System Prompts](/docs/run/agents/system-prompts)
* [Attaching Capabilities](/docs/run/agents/attaching-capabilities)