Choose a model
Different models suit different work. You choose which one answers you, and how much thinking it does first.
Pick the default model
Section titled “Pick the default model”Open Settings → General to pick the default chat model. The same page shows how that model reasons.
You can also change the model for a single session, from the field where you describe your task.
The models page lists every model running at Grunden today, what each one accepts, and what it costs.
Choose how much reasoning to spend
Section titled “Choose how much reasoning to spend”Some models let you set how long they think before they answer. More reasoning usually means a better answer and a slower one.
| Level | What it does |
|---|---|
| Minimal | Uses a very small amount of reasoning. |
| Low | Uses a lighter reasoning pass before answering. |
| Medium | Balances speed and reasoning depth. |
| High | Spends more time reasoning before answering. |
| Max | Uses the deepest reasoning the model offers. |
Not every model exposes these levels. Some decide their own reasoning path, and the app says so instead of offering a choice.
Check that your model can read your file
Section titled “Check that your model can read your file”You can attach files to a task, but not every model reads every type. If you attach something the current model cannot read, the app tells you before you send it.
| If you attach | And the model cannot read it |
|---|---|
| An image | Choose a model that accepts image input, or describe the image in words. |
| A PDF | Choose a model that accepts PDFs, or attach a Word file instead. |
| Audio or video | Choose a different model. |
The models page marks what each model accepts.