Thinking Modes
Arketic AI can think at different depths depending on how complex your question is. The thinking mode you choose tells the assistant how much reasoning to put into your next answer.
You set it from the brain icon in the chat input toolbar, and you can change it at any point in a conversation.
Where to Find It
The thinking mode control sits in the toolbar under the chat input box, next to the + (Add File) and Tools buttons.
It is a brain icon — not a text label, so it does not spell out the mode name on the button itself. Instead, the icon changes colour to tell you whether you are on the default or on a mode you picked yourself:
| How the icon looks | What it means |
|---|---|
| Grey brain icon | Auto is active — this is the default |
| Purple brain icon on a light purple background | A mode other than Auto is active (Fast, Deep or Ultra) |
Hover over the icon to see a tooltip with the current mode, for example Thinking: auto.
If the brain icon is purple, you are no longer on Auto. Open the menu and pick Auto to go back to the default.
The Four Modes
Clicking the brain icon opens a small menu with exactly four modes, in this order:
| Mode | What it's for |
|---|---|
| Auto | The default. Let Arketic decide how much thinking your question needs. |
| Fast | When you want a quick, direct answer. |
| Deep | When you want more thorough reasoning before the answer. |
| Ultra | When you want the most thorough reasoning available. |
The mode that is currently active is highlighted in purple in the menu.
Auto is the recommended setting for everyday use. You only need to change it when you specifically want a faster answer or deeper analysis.
How to Change the Thinking Mode
- Go to the chat input area at the bottom of the screen.
- Click the brain icon in the toolbar under the input box.
- Pick Auto, Fast, Deep or Ultra from the menu.
- The menu closes, and the icon turns purple if you chose anything other than Auto.
- Type your message and send it.
The mode you pick stays active for the rest of the conversation and is applied to every message you send afterwards. You can change it again between messages — for example, ask a quick question on Fast, then switch to Deep for a follow-up that needs more analysis.
Your choice stays in place until you change it again, so it is worth glancing at the brain icon before sending an important message.
Watching the Assistant Think
When the assistant produces reasoning before its answer, a small thinking bar appears above the reply.
While it is thinking:
- The bar shows a brain icon, the text "Thinking..." and three animated dots.
- It opens automatically, so you can read the reasoning as it is written.
- The reasoning appears in italic text in a scrollable box.
- Click the bar to collapse it if you would rather not watch.
When it has finished:
- The bar collapses on its own and shows how long the reasoning took — for example "Thought for 6s".
- Click the bar to expand it again and re-read the reasoning at any time.
- The final answer appears below the bar.
The reasoning text is read-only — you can't edit it or reply to it. It is there so you can follow and double-check how the assistant reached its answer.
The thinking bar appears whenever the assistant produces reasoning, so it is not tied to one specific mode. Deeper modes are simply more likely to show longer reasoning.
Choosing a Mode
There is no wrong choice — these are suggestions to help you decide:
| Situation | Suggested mode |
|---|---|
| Most everyday questions | Auto |
| "What does our policy say about...?" | Auto |
| "Summarise this document" | Auto |
| A short factual lookup you just want answered | Fast |
| "Compare these two options and recommend one" | Deep |
| "What are the risks of this decision?" | Deep |
| A large, multi-part analysis | Ultra |
| A detailed plan you want thought through end to end | Ultra |
Tips
- Start on Auto. Only switch when you have a reason to.
- Deeper modes take longer. Using Deep or Ultra for a simple lookup mostly means waiting longer for the same answer.
- Check the icon colour before you send. A purple brain icon means you are still on a mode you set earlier in the conversation.
- Read the reasoning when the answer matters. Expanding the thinking bar is a quick way to confirm the assistant understood your question the way you meant it.
- Thinking mode is separate from Tools. Web Search and Document Search are switched on from the Tools button — see Using AI Tools.
Related Pages
- Starting a Chat — the chat screen and its input toolbar
- Using AI Tools — web search, document search and action approvals
- Tips & Best Practices — getting better answers