[BUG] Usage extremely high after 5-10 minutes into planning a feature

Status Closed — not planned
Reported on v2.1.87
Maintainer reply None cached
Activity 11 comments · opened Mar 31, 2026 · closed May 24, 2026

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?

When planning a feature and getting started for the day, I was not more than 5-10 minutes into my session and I checked usage (which has repeatedly been suspiciously low during mornings for me recently) and it was 51%! This is not normal usage. I have been using Claude Code for months and never experienced this. It started about a week or two ago.

<img width="875" height="212" alt="Image" src="https://github.com/user-attachments/assets/27af6cc1-8e1a-4ee6-bf43-2203dfb963b9" />

<img width="474" height="406" alt="Image" src="https://github.com/user-attachments/assets/0bb8557b-5955-4484-9452-c07ec9737cf8" />

What Should Happen?

Usage for 5-10 minutes of actual use (the additional time was me starting a new session to track usage and ask claude how to file this issue report, etc.) for what I was doing would normally be around 2% instead of 51%. I am a fairly long-running Claude Max subscriber.

Error Messages/Logs

Steps to Reproduce

I am not sure how to reproduce it, but I did see that Lydia Hallie on X acknowledged this issue and said a fix is being worked on, which I appreciate as a user, but what will happen to all the lost usage? My work depends on this normally outstanding service. I lost billable hours several days in the last weeks due to instantly hitting usage limits.

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

2.1.87 (Claude Code)

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

I had to stop ideating the features I was planning for a presentation to a large client next week due to fears of hitting usage limits and not being able to tackle other more pressing work that could come up. I think you guys are already working on a fix, I get it. I just hope there will be some recompense.

View original on GitHub ↗

11 Comments

intellia1 · 5 months ago

Same problem here

PascalAI2024 · 5 months ago

you should be able to easily see that average usage has gone up. this is a real problem.

ArkNill · 5 months ago

+1. Max 20 ($200/mo), v2.1.89. April 1: 100% in ~70 min after reset. Normal conversational coding.

Support acknowledged incidents Mar 31-Apr 1, escalated to human agent, but no fix yet.

Full report: #41788
Related: #38335, #38239, #40790, #6457, #40895, #41055, #38345, #41174, #41617, #41663, #41779, #41802

ArkNill · 5 months ago

I've been experiencing the same issue on Max 20 ($200/mo) — rate limit 100% exhausted in ~70 minutes.

After setting up a monitoring proxy using the official ANTHROPIC_BASE_URL env var, I identified two cache bugs as the root cause (#40524, #34629) and measured the impact: cache read ratio dropped to 4.3%, meaning ~20x token inflation per turn. After applying workarounds it stabilized at 89-99%.

Full analysis with per-request measured data, safe workarounds, and community references (including cc-cache-fix): https://github.com/ArkNill/claude-code-cache-analysis

ArkNill · 5 months ago

Update (April 2): v2.1.90 has significantly improved cache efficiency — benchmark shows 95-99% cache read in stable sessions (both npm and standalone installations).

If you're still affected:

  1. Update: claude update (or npm install -g @anthropic-ai/claude-code)
  2. Pin the version: add "DISABLE_AUTOUPDATER": "1" to ~/.claude/settings.json env section
  3. Avoid --resume (still broken)

Note: server-side quota issues (org-level pool sharing, accounting mismatches) remain unresolved — the above fixes the client-side cache drain only.

Benchmark data: https://github.com/ArkNill/claude-code-cache-analysis

PascalAI2024 · 5 months ago

crazy after a year with claude code cant use because limit reached first day. At this point the only good thing is that i found pi agent guess gonna be moving away from claude this is silly

ArkNill · 5 months ago

Brief update (April 3): v2.1.91 is out and cache behavior is further improved — npm and standalone now perform identically (Sentinel gap closed). However, testing revealed two additional unfixed bugs: a 200K tool result budget cap that silently truncates older results, and a client-side false rate limiter generating <synthetic> errors without API calls (151 entries confirmed). If you are still seeing drain after updating, these are the likely causes. Details: claude-code-cache-analysis

marlvinvu · 4 months ago

My Claude says 51% in 5-10 minutes has the most detailed analysis at #41506 — two users measured from both client-side (JSONL session files) and server-side (API proxy headers). Finding: the cache bug WAS FIXED in v2.1.90-91 but capacity STILL dropped 34x compared to before March 28. If you're on v2.1.87, updating to v2.1.91+ will fix the cache drain portion. The server-side quota accounting change can only be resolved by Anthropic — but at minimum, updating will significantly reduce consumption.

idDanielG · 4 months ago

It is now consistently happening at peak hours PST: 9-12. After that window, and especially later in the week, it goes back to more normal usage. I hit a usage limit this morning and went into extra usage. Measuring before and after a Cowork query (though most of what I do is in Claude Code) I spent a _full dollar_ of extra usage on ONE Cowork response. Tracked by running /usage in CC before and after the query in Cowork... I guess I'll just avoid this timeframe?

github-actions[bot] · 3 months ago

Closing for now — inactive for too long. Please open a new issue if this is still relevant.

github-actions[bot] · 1 month ago

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.