Choose the model

Every reply's footer names the model that wrote it, so you can always see which one answered.

The model a group uses

A group runs on the first of these that is set:

  1. the group's own model, on its Overview page;
  2. the agent's model, on the organization page;
  3. the organization's default model;
  4. the platform default.

Effort (low, medium, high, xhigh, max) is chosen the same way. Group admins set a group's model and effort; Owners and Admins set the agent's and the organization's.

A change applies to each thread's next reply. The thread moves to a fresh session that reads the thread again, so nothing is lost.

Fast mode

Fast mode gives faster replies on Opus, at a premium price per token. It's per thread:

@cloud !fast
  • !fast alone switches it; !fast on and !fast off set it.
  • It applies to the thread you send it in. A !fork of that thread keeps it.
  • An Owner must turn on Allow fast mode on the organization page first. Until then the agent replies with where to turn it on.
  • The reply footer says when a reply was written in fast mode, and !status shows it too.

Switching models in a thread

Claude Tag lets you ask for a different model in a thread. Cloud Tag doesn't yet: the model is set per group, agent and organization, and !fast is the per-thread switch.