[BUG] Version 2.0.61 is extremely slow and causes rapid, excessive token consumption/rate limits
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?
After updating to claude-code version 2.0.61, the tool's performance has degraded significantly compared to version 2.0.56. The current version exhibits extremely high token usage for simple tasks, leading to the exhaustion of usage limits (rate limits) within minutes of starting a session (e.g., 15 minutes). Furthermore, the quality of the generated output ("export") is poor and often not usable, making the process inefficient and costly. This issue has rendered the current version practically unusable for my standard workflow.
These threads investigate recent Claude Code updates that allegedly decrease performance, increase token usage, and introduce new limitations:
What Should Happen?
Expected Behavior
I expect Claude Code to perform tasks with reasonable speed and token efficiency, similar to version 2.0.56. Usage limits should be hit only after substantial, high-quality work has been performed, not within minutes of a session start.
Current Behavior
Rapid Token Consumption: The tool burns through tokens at an unprecedented rate. My weekly/daily quota is depleted in approximately 15 minutes of use.
Poor Output Quality: The output (code, plan) often lacks quality and requires significant manual intervention, leading to wasted tokens on rework.
Performance Issues: The tool feels slow and unresponsive, negatively impacting the development workflow.
Usage Discrepancy: The reported token usage seems inflated compared to my actual activity and previous versions' consumption patterns.
Error Messages/Logs
Steps to Reproduce
⎿ Limit reached · resets 5pm (Asia/Dubai) · /upgrade to Max or turn on /extra-usage
/upgrade to increase your usage limit.
Claude Model
Opus
Is this a regression?
Yes, this worked in a previous version
Last Working Version
2.0.56
Claude Code Version
2.0.61
Platform
Anthropic API
Operating System
Windows
Terminal/Shell
Terminal.app (macOS)
Additional Information
_No response_
8 Comments
Found 3 possible duplicate issues:
This issue will be automatically closed as a duplicate in 3 days.
🤖 Generated with Claude Code
I'm not fully convinced this is strictly tied to version 2.0.61, but I’ve also been using Claude Code for several months and started noticing similar behavior on my side about a week ago. It burns through tokens extremely quickly now, and tasks that used to take around 10 seconds suddenly take several minutes (sometimes up to 8 minutes) to complete. For reference, I'm still on version 2.0.55, so the issue might not be limited to the latest release.
I have started experiencing this in the past two days. Using latest version as of now 2.0.65
Same here, after 2.0.69 today!
It's possible that your session had thinking on enabled by default. You can disable thinking via /config > Thinking mode: false in 2.0.70 and newer versions.
I have tried variations of switching models to sonnet and opus with thinking mode enabled and disabled but usage limits hits extremely fast. Just one small message takes around 6% of current session limit.
Usage has been excessive as of 2.0.6x and above, with or without /thinking activated. With the same type of prompts and workflow, my sessions have consistently reached their limits in the 4 hour window, where I previously (before 2.0.6x) only reached limits 20-30% of the time. With thinking activated, I could reach limits in less than an hour. The only other major change I made in this period related to workflow was giving claude skills.
This issue has been automatically locked since it was closed and has not had any activity for 7 days. If you're experiencing a similar issue, please file a new issue and reference this one if it's relevant.