Each subagent can pick its own model through the model field in frontmatter. The value is a model alias: sonnet, opus, or haiku. This lets you match the model to the difficulty of the work instead of running everything on one setting.
Match the model to the job
A rough guide:
- haiku for fast, simple, high-volume work: quick lookups, gathering file lists, straightforward summaries.
- sonnet for everyday coding, test writing, and most reviews.
- opus for hard reasoning: tricky architecture, subtle bugs, deep planning.
Running a light task on a fast model keeps things quick, while saving the strongest model for the jobs that actually need it.
Example
---
name: quick-summarizer
description: Summarizes a file or a diff in a few plain sentences.
tools: Read
model: haiku
---
You summarize. Read what you are given and return three or four
plain sentences. No preamble.
For a demanding agent, swap the alias:
---
name: architecture-reviewer
description: Reviews a design for scaling and coupling problems.
model: opus
---
You review architecture. Point out coupling, scaling, and
boundary problems, with concrete reasons.
Keep it flexible
If you leave model out, the agent uses the default. You can also change your main session's model any time with /model. Start most agents on sonnet, drop the cheap ones to haiku, and reserve opus for the few that genuinely reward it.
Comments
Be the first to comment.