Choosing the model (or Auto)
What the model picker is, what Auto does and when it makes sense to pick a specific model.
Cortex works with AI models from several providers. You can let it choose with Auto or pick one yourself.
Where it is
The picker is called Model and you'll find it in three places:
- On the create screen, next to the language, for both kinds of site.
- In the Studio Cortex bar, on an editable site.
- In the chat box of a Made by Cortex site.
Auto, the recommended option
Auto is selected by default. Cortex picks the best model for your idea and switches to another if one fails.
On a Made by Cortex site, Auto picks the model that codes sites best in Berse's tests. If one fails, another takes over.
If you have no reason to change it, keep Auto.
The models you can pick
| Provider | Models |
|---|---|
| Anthropic | Claude Sonnet 5.5, Claude Opus 5.5, Claude Haiku 5.5 |
| OpenAI | GPT-6 Sol, GPT-6 Astra, GPT-6 Luna, GPT-6.1 Sol |
| xAI | Grok 4.7, Grok 4.20, Grok 4.20 Reasoning |
| Gemini 4 Argon, Gemini 3.8 Flash, Gemini 3.5 Flash-Lite |
Each model has a label: Fast, Balanced or Top quality. In the compact list you see it when you hover over the model.
Does it change what I spend?
Credits reflect what each request actually uses. Top quality models usually use more credits than Fast ones. More on this in How credits work.
Berse remembers your choice
Your choice is saved in this browser. The one for editable sites applies to creating and to the Cortex bar. The one for Made by Cortex sites applies to creating and to their chat.
When to pick a specific one
- A result doesn't convince you and you want to try a different style of answer.
- You're making small changes and prefer a Fast model.
- A model doesn't respond: pick another and try again.