Overview
Julius lets you choose which AI model powers your analysis. Use the Model Selector in the chat dropdown to switch between models depending on your task. Models are split into two groups: Julius models (tuned for data analysis) and General models (third-party models from OpenAI and Anthropic).Julius models
Julius models are purpose-built for data work. They’re the default and the best starting point for most tasks.| Model | Available on | Best for |
|---|---|---|
| Julius 1.2 Lite | All plans | Quick questions and simple lookups |
| Julius 1.2 | Paid plans | Most questions and everyday analysis |
| Julius 1.2 Max | Paid plans | Complex, multi-step analysis |
| Julius 1.2 Ultra | Max plan and above | The most demanding, multi-step work |
General models
You can also use third-party models directly. These are useful when you want a different reasoning style or want to compare outputs.| Model | Available on | Description |
|---|---|---|
| GPT-5.5 | Paid plans | OpenAI’s most capable model |
| Claude Opus 4.8 | Paid plans | Anthropic’s most powerful model |
| Claude Sonnet 4.6 | Paid plans | Anthropic’s fast, balanced model |
Plan availability can change. Check your workspace settings for the most current details.
Code environment
The model selector also includes your code environment settings:- Sandbox is the container where Julius runs code. You can reset, delete, or expand resources (RAM/CPU) from here.
- Code runtime lets you switch between Python, R, Lean Python (a lighter Python environment), and GPU.
Choosing the right model
- Start with Julius 1.2 unless you have a specific reason to switch. It’s tuned for data analysis and handles most tasks well.
- Use Julius 1.2 Max (or Julius 1.2 Ultra on higher plans) for complex prompts that involve multiple steps or large datasets.
- Use Julius 1.2 Lite when you need a fast answer to a simple question.
- Try a General model for a different reasoning style, or to compare outputs across providers.