Skip to main content

Guide 03 · Claude edition · Free to read online

Claude Models & Reasoning Guide

Choose the right settings for the work.

Start with a capable model, adjust the effort to the assignment, and keep an eye on your usage. Here are the settings from the October 2026 guide.

Get all three free PDFs

01 · Starting settings

Start with Opus 5.5 on Max.

Higher effort means more thorough responses, but takes longer and uses your limits faster.

Model · this session

  • Opus 5.5Selected
  • Sonnet 5.5
  • Fable 5.1
  • Haiku 5.5

More models · Fast mode

Effort

  • Low
  • Medium App recommendation
  • High
  • Extra
  • Max3.5× or more usage
  • Ultracode

These model names and controls are retained from the October 2026 guide. Check your account for the options currently available to you.

02 · More demanding work

When to go bigger.

Fable 5.1
The answer depends on connecting facts across a whole file at once, like a project recommendation built from the full record.
Ultracode
The same prompt across many documents, like the same question in 40 reports. In the Code tab, choose Effort → Ultracode. Switch back to Max when done.

03 · Manage your usage

When to scale back.

If Settings → Usage doesn’t say On track, switch to:

Sonnet 5.5 · High
Jobs under 10 pages in and 2 pages out, like emails and short summaries.
Opus 5.5 · High
Everything bigger.

What to check in Usage

Look at both the current session and the week, along with their reset times. The guide’s example shows On track, with 21% of the current session and 15% of the week used.

Make the assignment count, whatever model you choose: read the prompting guide.

Keep a copy

Three guides.
One useful reference set.

Save them to your files, print them for your desk, or keep them handy during your next assignment.

  1. 01Prompting Guide
  2. 02Plugins & Connectors Guide
  3. 03Model & Reasoning Cheat Sheet

All three are free to read online. Sign up with your email to download the PDF bundle. These Claude guides use everyday work examples.

Work with Wesley

Connect it. Configure it. Put it to work.

The guides explain the ideas. In the $500 setup, we connect your approved data and software, configure your AI, and practice a real task together. You leave with tested connections and instructions you can use again.