Setup review by Anthropic’s model guide
Explore burning big-model quota on every file read: wrong-model subagents and ill-fitting effort found, a config fix each.
3.4K saves on X, as of
Read this guide: https://claude.dev/blog/which-claude-model-should-you-build-with
Then review my Claude Code setup against it: my user and project settings, the subagents in ~/.claude/agents and .claude/agents, and CLAUDE.md. Find three kinds of problems:
1. Subagents on the wrong model. Subagents that only search files, read logs or run tests but set no model, so they run on my main model. If I haven't defined my own subagent named Explore with model: haiku, count that too. The other way around, subagents that edit code or make judgment calls should stay on the main model, so check whether CLAUDE_CODE_SUBAGENT_MODEL is pushing them down to a small model too (check only that one variable, in the env block of my settings and in my shell).
2. A model or effort level that doesn't fit the work. Use the guide's table: Opus 5.5 at medium for everyday engineering and debugging, Sonnet 5.5 for well-scoped work like a bug fix or a feature built against a spec, and Fable 5.2 for major refactors, migrations and leading many subagents. Flag a default effort above high, and suggest testing the next model up at a lower effort first. Flag opusplan as my default too, because it puts code edits on Sonnet.
3. No way for you to verify your own work in this project: CLAUDE.md doesn't say how to run the tests or the build, or there's no check to run at all. The guide's order is to make sure you can verify your work first, then raise effort one level, and only then move to the next model up.
For each issue, quote my setting and the matching line from the guide, explain what it costs and where it will go wrong, and suggest the smallest change, written as config I can paste.
Then look at this project's code and recent commits and give me a table: which kinds of work here fit which model and which effort level.
Last, tell me how to use /model and /usage to run one task from my backlog on two models and compare the cost.
Show me the recommendations first. Don't change anything yet.Use it
Send it to Fable 5.2 in Claude Code.
You get Each issue with your setting and a config fix to paste, then which work fits which model and effort.
Anthropic just published its full guide to choosing between Fable 5.2, Opus 5.5, Sonnet and Haiku. You can ask Fable 5.2 to use it to review your Claude Code setup.
Claude Code’s built-in Explore subagent used to always run on Haiku. Now it runs on your main model. If that’s Opus 5.5, or you just switched to Fable 5.2, Explore is burning big-model quota for nothing every time it reads files or searches code. So define your own subagent named Explore and set model: haiku.
Send this prompt to Fable 5.2.
Add Copybook to your agent, then ask it next time. Set it up
