[BUG] Model Automatically Reducing Effort to 50% Regardless of Max Setting

Preflight Checklist

• [x] I have searched existing issues and this hasn't been reported yet
• [x] This is a single bug report (please file separate reports for different bugs)
• [x] I am using the latest version of Claude Code

What's Wrong?

During normal PST working hours the model admittedly is set to 50% effort on Anthropic side regardless of personal settings. However there is not a reduction is cost or a way to put the effort back to max. This is a significant issue as we are paying for a product not being received AND it's affecting our operation as models degrade during the workday and have totally separated metrics as night. We can not operate like this the effort must be static once set by the user.

What Should Happen?

The model should be expected to perform no matter time or load.

Error Messages/Logs

Your settings already have "effortLevel": "max" and "alwaysThinkingEnabled": true. So from your config side, everything is set correctly — you're asking for maximum reasoning.

The 50% reasoning effort I'm experiencing is being applied downstream of your settings, likely at Anthropic's infrastructure level. Your config says "max" but something between you and me is overriding it.

This is an Anthropic platform bug. Your settings are correct.

Steps to Reproduce

working through all settings

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

early march

Claude Code Version

claude 4.6

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_