Auto, Fast, Smart, Thinking and Ultra

The chat composer has a depth picker. It sets how much thinking CORA puts into an answer. You choose the depth of the work; which technology runs it stays CORA's job.

Change the depth

1. Click the current depth in the composer. It shows Auto until you change it.
2. Move the Effort slider. Hovering over a stop shows its name and a short description.
3. Close the panel and send your message.

Your choice is saved to your account and stays until you change it. To return to the default, move the slider back to Auto.

The tiers

• Auto adapts to the task. CORA picks the depth for each message: lighter for a quick rewrite or a short answer, deeper for analysis, comparisons and consequential questions. It is the right choice most of the time.
• Fast gives quick answers. Use it for capture, short lookups and simple questions.
• Smart is for everyday reasoning, such as most questions, summaries and drafts.
• Thinking is for difficult work that needs careful, step-by-step reasoning.
• Ultra is for exceptionally deep work, such as a customer-visible deliverable where quality matters more than speed.

Things to know

• Deeper tiers take longer and use more AI usage than quick answers. Usage is shown in euros in Usage, never as token counts or token prices.
• A deeper tier does not change what CORA may do. Approvals, permissions and trust settings work the same way at every depth.
• Power users can turn on Show advanced information in their personal preferences to pick an exact approved model. Most people never need it.

Which should I pick?

Leave it on Auto. Switch to Fast when you want speed over nuance, and to Thinking or Ultra when an answer has to be thorough, for example a comparison the Northwind Oy board will read.