Efforts

An agent is a backend, a model and an effort: how hard to think. Set one when a task needs more or less thought than the agent's default.

claude / claude-opus-4-8 : high
  │           │            └── effort
  │           └── model
  └── backend

The word belongs to each backend rather than to humanize, so the values differ.

Try it

hmz exec -f ralph_loop -a claude/claude-opus-4-8:high "fix the build"

The effort row of the agent's sheet shows high. Press ←/→ to adjust it; the swarm row turns swarm mode on for a model that has one.

Set the effort

backend/model:effort is the short spelling. The written-out form of -a takes the same thing, and so does a flow's Python config:

command line

hmz exec -f ralph_loop -a cli=kimi,model=kimi-code/k3,effort=swarmmax "fix the build"

Python

ClaudeCodeAgentConfig(model="claude-opus-4-8", effort="high")

Efforts by backend

humanize does not check an effort against a list: a value your account has but this page does not still works. These are the backends whose ladders need explaining; the whole set is in Agents › Efforts.

Backend	Efforts, hardest first
Claude Code	ultracode, max, xhigh, high, medium, low
Codex	ultra, max, xhigh, high, medium, low — each model takes its own subset
Kimi Code	max, high, medium, low, each also as swarm…
pi	max, xhigh, high, medium, low, minimal, off
opencode, mimocode	the model variant: xhigh, high, medium, low, minimal
ZCode	max, high, low, enabled, nothink, disabled — two vocabularies, and a model takes one of them

• ultracode is Claude Code's xhigh thinking with the turn opted into orchestrating a fleet of its own, so it sits above max. It is real and undocumented, and no listing the CLI answers with will ever name it. humanize keeps it anyway.
• Kimi Code's effort says how wide as well as how hard. max is one agent; swarmmax is the same thinking at the width of a fleet of subagents. The prefix is exported as hmz.flows.SWARM, which is where a flow steering by it reads it.
• pi's off is the model asked not to think at all — the least of the efforts, not the absence of a setting.
• Codex's models differ from each other. gpt-5.6-sol takes ultra; gpt-5.5 does not, so the interface offers each model only the efforts it takes.
• ZCode's ladder is two vocabularies in one. The models that take a thinking budget answer max, high and low, with nothink at the bottom; the ones that only take thinking-or-not answer enabled and disabled. Each model is offered the rungs it said it takes, and no model takes both halves.

Change the effort while the flow runs

The rest of this page is the weaver's — whoever wrote the flow.

A config is frozen. A session resumes under the settings it opened with, and a config that changed mid-flow would silently split one conversation across two models. The effort is the one setting a flow may move as it goes:

agents.builder.effort = "low"       # every session of this agent, from its next turn
session.effort = "max"              # this conversation alone
session.effort = ""                 # and back to whatever the agent runs at

A swarm prefix moves with it: agent.effort = "swarmmax". Read it back through the same property:

agent.config.effort	what the agent was configured with
agent.effort	what its turns actually run at

The change takes hold on the next turn. The turn already under way keeps the effort it started at: a model does not think harder halfway through an answer. How it lands is the backend's own business:

Backend	How the new effort reaches the model
Codex, Kimi Code, opencode, mimocode	sent with each turn
Claude Code	an argument of the process it is held open as, so that process ends and the conversation resumes in one started at the new effort
pi	it has a command for it, and is told
ZCode	its app server keeps the level on the session, and is told the new one before the next turn

What to steer by

The reading that responds to effort is juice(): output tokens an average turn of the model came out with. A model asked to think harder writes more in each answer, so that average is what an effort moves.

agent.juice(over=60)

official/fixed_juice_ralph governs on it — a Ralph loop that moves the effort a rung a round to hold the agent to a target.

See also

• Cost and rate
• Permissions — the other thing set on the model sheet
• Agents › Efforts