AI Chat model configuration
Each chat uses one model from the Syntic Amara family, chosen in the model menu above the message box. If you never change it, chat uses your plan’s default, which is Amara S1.2. The same menu turns Adaptive thinking and Web search on or off.
Change the model
Open the model menu
Click the model’s name beside the message box.
Pick a model
Each entry shows the model name and a one-line description. A model your plan doesn’t include has a lock icon; choosing it shows which plan it needs, for example requires the Pro plan or higher.
The choice is saved with the conversation, so going back to an older chat keeps the model you used there.
Adaptive thinking
Adaptive thinking lets the model think longer when a request is complex. When it is on, the menu label shows Adaptive after the model name. Toggle it in the same menu or with ⇧⌘E (Shift+Ctrl+E on Windows). Longer thinking uses more tokens.
Web search
The Web search switch decides whether the assistant may look things up online. With it off, it can’t search at all. If the switch is greyed out, your organization or your role doesn’t allow web search. Searches count toward your plan’s daily allowance; see Daily limits on searches and uploads.