[BUG] approaching usage limit

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?

<img width="1920" height="1080" alt="Image" src="https://github.com/user-attachments/assets/dd044a9e-17ad-4a08-b84e-9da1e9952107" />

Subject:** False "Approaching usage limit" warning persists at 51-52% weekly usage on Max plan — chat support (Fin) unable to resolve, requesting human escalation

---

Account / Plan: Max plan, Claude Desktop app on macOS
User: Vardhman Mehta (vardhmanmehta@gmail.com)
Date observed: 2026-05-04, persistent through evening

Issue:
The "Approaching usage limit" warning is showing in the chat composer area even though my actual weekly usage is only 51-52%, with 4 days remaining until reset. The warning persists across:

• Restarting the Claude desktop app
• Full logout and re-login
• Starting new conversations
• Switching between sessions

Prior support attempt:
I have already engaged with the Anthropic chat support (Fin AI agent) on this issue — the AI agent was unable to identify the root cause or provide a resolution. Requesting escalation to a human support engineer who can investigate at the platform level.

Expected behavior:
"Approaching usage limit" should only appear when actual usage is genuinely close to the hard limit (e.g. 90%+), not at the 50% mark with 4 days of reset window remaining.

Actual behavior shown in screenshots:

• Screenshot 1: Weekly · all models at 51%, all other meters at 0%, "Approaching usage limit" warning visible
• Screenshot 2: Weekly · all models at 52%, 5-hour at 3%, all other meters at 0%, same warning still visible

My suspicion:
The warning text appears tied to either (a) the context-window indicator at the top showing 510.4k / 1.0M (51%), and the UI is conflating "context window getting long for this conversation" with "user is approaching their account quota," OR (b) the warning threshold is hard-coded at 50% weekly which is far too conservative for a 7-day rolling window.

Either way, this creates user anxiety about hitting limits when no real limit is being approached. On a paid Max plan, this is a meaningful UX issue — I've stopped working multiple times today thinking I needed to throttle.

What I've tried (in order):

1. Engaged Anthropic chat support / Fin AI agent — no resolution provided
2. Quit and re-launched Claude desktop app — warning persists
3. Logged out and logged back in to my account — warning persists
4. Started fresh conversations — warning reappears as soon as context grows
5. Cleared older chats — no effect on the warning

None of these cleared the warning. It reappears in every new conversation as soon as the context window crosses ~50%.

Request:

1. Please escalate to a human support engineer (chat AI agent has been unable to assist).
2. Please confirm whether this warning is tied to context-window usage (not weekly quota) — if so, please consider re-wording to avoid implying a billing/quota issue.
3. If it IS triggering on weekly quota at 50%, please consider raising the threshold (90% would be reasonable for a "approaching limit" warning).
4. Please advise if there is a setting to suppress this warning, or whether this is a known issue being addressed.

Attachments: 2 screenshots (timestamps 22:31 and 22:55 on 2026-05-04, both showing 51-52% weekly usage with the warning visible).

Impact: Workflow disruption — false-positive limit warnings cause unnecessary work-throttling on a paid plan that has substantial remaining capacity. As a Max-tier subscriber I expect support pathways that resolve issues the chat AI cannot — this ticket is requesting that human pathway.

Thank you.

— Vardhman

.

What Should Happen?

the approaching usage limit error must not appear as I have enough weekly and daily limit

Error Messages/Logs

Steps to Reproduce

<img width="1920" height="1080" alt="Image" src="https://github.com/user-attachments/assets/3c12f1d0-1f9c-476a-a412-aa2570d6c0c2" />

Claude Model

Opus

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

1.5354.0

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_