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

# Xavier

> Xavier (xavier-r1) is our specialized 70B parameter reasoning model built on top of Llama 3.1. It excels at emotional reasoning and complex problem-solving, making it ideal for applications requiring nuanced understanding and thoughtful analysis. This model is currently under limited release.

## Key specifications

| Specification     | Value                      |
| ----------------- | -------------------------- |
| Parameters        | 70 billion                 |
| Base architecture | Llama 3.1                  |
| Context window    | 128,000 tokens             |
| Input format      | Text                       |
| Output format     | Text with reasoning tokens |
| Release status    | Limited release            |

## Distinctive features

Xavier has several unique capabilities that set it apart:

* **Thinking tokens**: Outputs explicit reasoning tokens that show the model's step-by-step thinking process
* **Emotional reasoning**: Particularly strong at understanding and reasoning about emotional contexts
* **Multi-step planning**: Excels at breaking down complex problems into manageable steps
* **Deliberate processing**: Takes more time to generate responses but produces more thoughtful results

## Use cases

Xavier is particularly well-suited for:

* Mental health support applications
* Complex decision support systems
* Educational tools requiring step-by-step explanations
* Research assistance requiring nuanced reasoning
* Relationship coaching and conflict resolution
* Content that requires emotional intelligence and sensitivity

## Example usage

```javascript theme={null}
import OpenAI from "openai";

const openai = new OpenAI({
  baseURL: "https://api.animusai.co/v2",
  apiKey: process.env.ANIMUS_API_KEY,
});

const completion = await openai.chat.completions.create({
  model: "xavier-r1",
  messages: [
    { 
      role: "system", 
      content: "You are Xavier, an assistant that shows step-by-step reasoning and has strong emotional intelligence." 
    },
    {
      role: "user",
      content: "I'm having a conflict with my colleague. They keep interrupting me in meetings, and I'm getting frustrated. How should I approach this situation?",
    },
  ],
});

console.log(completion.choices[0].message);
```

## Performance considerations

Xavier's unique thinking capabilities come with some important performance considerations:

* **Increased token usage**: The explicit reasoning process uses more tokens than standard responses
* **Longer generation time**: Takes more time to generate responses due to the additional reasoning steps
* **Higher resource utilization**: May require more computational resources for optimal performance

## Access and availability

Xavier is currently under limited release. To request access to Xavier, please reach out directly to the Animus team at [support@animusai.co](mailto:support@animusai.co) with details about your use case.

## Best practices

To get the most out of Xavier:

* **Be specific**: Provide clear instructions about the reasoning you want to see
* **Allow sufficient time**: Expect and plan for slightly longer response times
* **Consider context limits**: Be mindful of the 8,000 token context window, especially with reasoning tokens
* **Optimize token usage**: For simpler tasks, consider using Vivian instead to save on token costs

For more information about leveraging Xavier's reasoning capabilities, see our [Text Generation](/essentials/text-generation) guide.