[BUG] Sonnet output quality decline over several weeks

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?

The Sonnet model has been getting progressively worse over the last few weeks. I have noticed this across projects. I briefly switched to Opus, but the token rate is so high that it doesn't warrant the operational cost.

What Should Happen?

Consistent output quality. I have worked with Claude for over 12 months personally and at work. The Sonnet model is outputting lower quality work than before and taking more time to resolve a technical issue.

I would expect the opposite behavior as LOTR is used to improve the Model. It is beginning to feel like a soft push to use Opus. But the $$$ to quality costs are prohibitive for an enterprise agreement.

Error Messages/Logs

Steps to Reproduce

This is an overall quality reduction and not specific to a single code pattern.

Claude Model

Sonnet (default)

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

Claude Code CLI: 2.1.252

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

VS Code integrated terminal

Additional Information

The reasoning quality has dropped for some very basic coding. With more frequent coding bugs.