Claude gives you two dials before you start: which model to use, and how much effort it should spend. The default is fine for most tasks. Knowing when to change it saves time on easy work and improves results on hard work.
The model picker
The model picker currently shows four models. They differ in speed, depth, and how much of your plan's usage they consume. Names and lineups change, so read the picker's own descriptions instead of relying on a list from last month.
Two things to check:
- Plan availability. Not every model is included on every plan.
- Usage credits. On certain plans, some models require usage credits beyond what your plan includes.
The Effort setting
Effort controls how much work Claude puts into a response. You will find it in the same menu as the model name, usually set to Medium, which is marked Recommended. The levels are Low, Medium, High, Extra, and Max. Higher effort means a more thorough result, a slower response, and more of your usage. Max carries a warning that it can use four times as much or more. Lower effort is faster and lighter.
Matching the dial to the task
| Task | Starting point |
|---|---|
| Quick lookup, a short rewrite, a simple format change | Lower effort, any model that responds fast |
| Everyday drafting, summaries, and email replies | Medium effort |
| Analysis, comparisons, and careful editing | High effort |
| Long multi-step jobs, complex reasoning, or work with real consequences | Higher effort on a stronger model |
Treat the table as a starting point, and adjust from what you see.
A way to decide
Ask what a mistake would cost. A slightly weak answer to a throwaway question costs nothing, so choose speed. A weak answer in a client proposal, a financial summary, or a plan others will act on costs a lot, so choose depth and check the result.
Watch the cost
Higher effort and stronger models use more of your plan. Make a habit of two checks: whether the task needs the stronger setting, and whether you can get most of the benefit by raising effort on one step instead of the whole job. If you reach your limit often, review which tasks truly need the top settings.
Review, whatever you pick
No setting removes the need to review. The Effort setting changes how much work goes into an answer, not whether the answer is correct. Check facts, numbers, and anything others will rely on.
Go deeper
Module 1 of Claude Unlocked walks through the current model picker, with a short exercise to try, and Module 3 explains the Effort setting. The guide includes cost notes, because not every model is included on every plan. It has 78 prompts across 10 modules, is updated for the October 2026 interface, and costs $39.99, one time. The complete list of what changed is on the What's New page near the front of the guide.
That completes this series on what changed in Claude Unlocked v3.1.
Published by RightHand AI. Drafted with AI assistance and reviewed and finalized by RightHand AI before publication.
Claude Unlocked is independent training material about Claude, published by RightHand AI. It is not an official Anthropic product.
RightHand AI, a DBA of Ussen Technologies, LLC, 5233 Bellaire Blvd., Suite 638, Bellaire, TX 77401.