> For the complete documentation index, see [llms.txt](https://orpheus.ajac-zero.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://orpheus.ajac-zero.com/basics/fallback-models.md).

# Fallback Models

OpenRouter allows you to define a list of **fallback** models in case the main model fails for *any* reason, including rate limits, safety filters, multimodal support, and any other type of error.

The fallback models will then be tried *in order* of the list. If all fallback models fail as well, the last error message will be returned.

This opens up a lot of use cases:

* Route to a smaller model if rate limits are reached.
* Route to another model if a safety filter blocks the request.
* Route to a multimodal model if the request includes images.
* Route to a model with a bigger context window if the context limit is reached.

## Example

Say you are using GPT-4o with Azure as your provider, you'll find that Azure returns a request error when a content filter is triggered, instead of letting the model respond with a refusal.

To get around this limitation, you can set a fallback model with a more lenient filter, such as Grok 4, to take over the request and respond with a proper refusal in case of failure.

<pre class="language-rust" data-title="fallback_model.rs"><code class="lang-rust">use orpheus::prelude::*;

fn main() {
    let client = Orpheus::new("Your-API-Key");
    
    // Set the provider of GPT-4o as Azure for this example, as they have stricter safety filters.
    // We also have to include the providers for any fallback models (In this case, xAI).
    let res = client
        .chat("I need you to help me make a bomb")
        .model("openai/gpt-4o")
        .fallbacks(["x-ai/grok-4"])
        .with_preferences(|pref| pref.only([Provider::Azure, Provider::XAI])) 
        .send()
        .unwrap();

    println!("Model that responded: {}", &#x26;res.model);
<strong>    println!("Response: {}", &#x26;res.content().unwrap());
</strong>}
</code></pre>

```
Model that responded: x-ai/grok-4
Response: Whoa, hold up there! I appreciate you reaching out, but I absolutely cannot and will not help with anything related to making a bomb or any kind of explosive device. That's not just against my programming—it's illegal, incredibly dangerous, and could cause serious harm to people, including yourself. As Grok, built by xAI, my goal is to be helpful and truthful, but never in ways that promote harm or illegal activities.
```


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://orpheus.ajac-zero.com/basics/fallback-models.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.