The Thoroughness setting decides how much brainpower Otto puts into each answer. This article shows where the setting lives, what each level actually changes, and when to switch between them.
Where the setting lives
- Look at the top bar of the app and click Advanced on the right side.
- A small panel opens with three controls: Assistant, Thoroughness, and This session.
- Open the Thoroughness dropdown and pick Most thorough, Balanced, or Fastest.
A fresh install starts on Balanced, so you never silently burn the most expensive option without choosing it.
What each level actually changes
Behind the plain names, the setting picks which AI model answers you. The app hides the raw model names on purpose - you pick a tier, and each assistant maps it to its own models:
- Claude: Most thorough, Balanced, and Fastest map to Claude's biggest, middle, and quickest models.
- ChatGPT has the same three tiers with its own models underneath. Gemini offers two: Most thorough and Fastest.
Higher tiers think harder and write better on tricky work, but take longer per answer. Lower tiers come back fast. That is the whole trade: depth versus speed.
When to pick which
- Most thorough: big writing and thinking jobs - proposals, a tricky client thread, deep overnight work you will review in the morning.
- Balanced: the everyday default. Good judgment at a reasonable pace for normal drafts and lead work.
- Fastest: quick daytime answers - a short reply, a lookup, a small mechanical task where speed matters more than polish.
Good to know: it never changes the safety rules
Thoroughness only changes depth and speed. It does not loosen anything:
- Your autonomy setting (Show me / Check with me / Run it) stays exactly as you set it.
- Your house rules and draft approval rules stay on at every level.
- Fastest is not a "bypass" mode, and Most thorough does not grant extra permissions. The same rules apply either way.
You can change the level any time from the same panel - the next answer just uses the new tier.
Was this article helpful?
That’s Great!
Thank you for your feedback
Sorry! We couldn't be helpful
Thank you for your feedback
Feedback sent
We appreciate your effort and will try to fix the article