When you see "Claude usage limit reached"
First work out whether you hit the rolling 5-hour window or the weekly cap — they recover in completely different ways. Here is how to tell, and what to do when you are out.
Four things to know
Rolling window
A short-cycle cap on a rolling window. It restores on its own — nobody needs to be contacted.
Weekly cap
A longer-cycle total. Once you hit it, waiting out the 5-hour window will not help.
Temporary official boost
On 2026-08-18 Anthropic said the 50% weekly boost is extended through 2026-08-31.
Two different reset times
On the QCode side: the daily limit resets at 00:00 Beijing time; booster quota resets separately at 02:00 the next day.
What the message actually says
The full string is usually "Claude usage limit reached. Your limit will reset at …" followed by a specific time. That time is the tell: a few hours out generally means the rolling 5-hour window; a few days out means the weekly cap. Both are set by Anthropic per subscription tier and have nothing to do with your network, client version or API key.
The current temporary measure
On 2026-08-18 Anthropic announced that the 50% weekly allowance boost is extended through 2026-08-31, while noting capacity pressure. It is a temporary measure; there is no official statement about what happens after Aug 31. Do not plan long-term capacity around it.
Timeline
Many developers publicly log "most of my allowance gone in minutes", usually alongside parallel threads and whole-repo sweeps.
Anthropic extends the 50% weekly boost through Aug 31 and mentions capacity pressure.
The error string itself stays stable and remains the fastest way to tell the two caps apart.
Work out which one you hit first
What is confirmed
The string "Claude usage limit reached. Your limit will reset at …" appears consistently in public discussion; on 2026-08-18 Anthropic announced the 50% weekly boost is extended through Aug 31.
No evidence — do not treat as fact
Some claim that under load the vendor silently routes requests to a smaller or quantised model. We searched public discussion and found no first-person evidence at scale and no official statement, so this page does not make that claim. Being refused, truncated or blocked by a classifier is a different phenomenon from "they swapped the model".
The two caps
Rolling 5-hour window
A short-cycle ceiling. It restores automatically at the time printed in the error. You normally hit it through dense use in a short span.
Weekly cap
A longer-cycle total. After hitting it, waiting for the 5-hour window does nothing — you either wait for the cycle to reset or continue on another path.
Why some people burn it in minutes
Public reports share a few patterns: running several threads in parallel, letting the model sweep an entire repository, and spending the most expensive model on work a small one would finish. One developer logged 23% of the allowance gone in six minutes; another hit the 5-hour cap in twenty minutes with four threads. Tightening scope, forbidding large refactors outright and pushing simple work to cheaper models are the fixes the community keeps confirming.
What QCode can and cannot do here
What it can do: when you are out, switch model family and keep working — one key covers Claude, GPT, Gemini, GLM, Kimi, DeepSeek and Qwen, so you do not need a second subscription. If today’s allowance runs short you can also self-serve a booster pack. What it cannot do: if you are still calling Anthropic’s own line, their 5-hour window and weekly cap still apply. We cannot remove them.
FAQ
How do I tell which cap I hit?
Read the reset time in the error. A few hours out is usually the rolling 5-hour window; a few days out is the weekly cap. It is the fastest and most reliable signal.
Do I need support to lift the 5-hour window?
No — it restores on its own. Treat any offer to "instantly lift an official cap" with suspicion.
Is the official 50% boost still on?
Per the 2026-08-18 announcement it runs through Aug 31. Nothing official has been said about after that, so do not plan long-term capacity around it.
Why does one small fix eat a big slice of the weekly cap?
Usually because the model went out of scope: it swept the repo, questioned the architecture, refactored config, and only then touched the bug. Pinning the scope and forbidding large refactors normally helps a lot.
Does QCode have a 5-hour window too?
No. Personal plans dropped the 4-hour window entirely in July 2026; the only dimension now is a daily limit that resets at 00:00 Beijing time. If it runs short you can self-serve a booster pack, which resets separately at 02:00 the next day — deliberately two hours later than the daily limit, leaving a buffer for work that runs past midnight.
Will switching to QCode make this error go away?
If you are still calling Anthropic’s own line, their caps still apply. The value is that when you run out you have other model families to switch to immediately — not that the official cap disappears.
Sources
The error strings and announcement dates cited here come from publicly searchable developer discussion and Anthropic official posts between 2026-07-30 and 2026-08-20.
Out of quota? Six other model families are still there
One key across seven model families; the daily limit resets at 00:00 Beijing time and boosters are self-serve.
Related reading
How Claude Code quota resets work
Cycles, reset times and common misconceptions.
Claude Code cost optimisation
Concrete ways to spend the allowance where it counts.
QCode plans and billing
Daily limits, booster packs and multi-month discounts.
The official cap mechanics and temporary boost described here follow Anthropic announcements and may change at any time. QCode-side reset behaviour follows this site’s pricing page.