[BUG] Overthinking
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?
Shocking amount of token consumption during standard thinking in 2.0.76. Performance dramatically slower -- simple actions taking 1-2 minutes. Consumption way higher. Getting 15k tokens of thinking without tool use on one to two sentence request, and no suggestion of ultrathink --> altering 10 lines of code.
What Should Happen?
Normal behavior.
Error Messages/Logs
Steps to Reproduce
Make any old request.
Claude Model
Opus
Is this a regression?
Yes, this worked in a previous version
Last Working Version
2.0.75
Claude Code Version
2.0.76
Platform
Anthropic API
Operating System
macOS
Terminal/Shell
Terminal.app (macOS)
Additional Information
No response
Found 3 possible duplicate issues:
- https://github.com/anthropics/claude-code/issues/3843
- https://github.com/anthropics/claude-code/issues/13173
- https://github.com/anthropics/claude-code/issues/14129
This issue will be automatically closed as a duplicate in 3 days.
- If your issue is a duplicate, please close it and 👍 the existing issue instead
- To prevent auto-closure, add a comment or 👎 this comment
🤖 Generated with Claude Code
I am using -output-format stream-json --verbose.
After connecting, when I instruct Claude Code to reference image file paths (such as .jpg files) located in a specific folder on my system, Claude Code immediately terminates the CLI session as soon as a tool is invoked (it crashes / force-closes).
This happens consistently on other computers as well.
It really feels like a strange bug in Claude Code. The behavior is very abnormal.
I am located in Korea, and this issue occurs in a Windows environment.
My suspicion is that this might be related to language or character encoding issues (for example, non-ASCII paths, Korean characters, or Windows encoding).
Please investigate this as a potential bug and help resolve the issue.
This issue has been inactive for 30 days. If the issue is still occurring, please comment to let us know. Otherwise, this issue will be automatically closed in 30 days for housekeeping purposes.
Closing for now — inactive for too long. Please open a new issue if this is still relevant.