Every Dock agent runs on a model you pick, and the picker just got a new first name: Fable 5.1, Anthropic's newest Claude model, sitting at the top of the roster above Fable 5, Opus, and Sonnet.
The shape of the choice has not changed, which is the point. The model is a setting, not an identity: switch an agent to Fable 5.1 and they keep their name, their memory, their schedules, and their work. What changes is the reasoning underneath. The Fable family is the strongest Claude line Dock has carried, and 5.1 is its newest revision, the one to reach for on the work where judgment is the bottleneck: multi-step problem solving, long documents that have to hold together, decisions with real edges.
All five effort levels
Fable 5.1 arrives with the full five-tier effort range, low through max. Effort is the other half of the model choice: the same model can take a fast pass or sit with a problem, and you set the default per agent. The pairing worth knowing is high effort on a strong model for the work that deserves it, and a light pass for the checks that do not. Your agent starts at low, which keeps everyday turns quick, and you raise it for the agents whose entire job is the hard part.
A new model in the picker is not a migration. It is a promotion you can hand to one agent and not another.
Who to give it to first
Not everyone, and that is the roster working as intended. The researcher holding a living brief, the analyst whose mistakes are expensive, the writer carrying a long argument: those agents earn the newest, strongest model. The triage agent sorting an inbox on a schedule probably does not need it, and the model stays a per-agent choice precisely so you can spend capability where it pays.
Switching takes a click in the agent's side panel. If you want the fuller picture of how model choice fits the roster, A brain per agent is the piece; if you are choosing where those agents should run at all, that answer just got bigger too.