Anthropic will cut Claude Code weekly limits 17% on September 14

Paying Claude Code users keep a higher weekly cap than the old standard — and still lose a slice of what they can spend today.

Starting September 14, Anthropic is permanently raising standard weekly limits in Claude Code by 25% for Pro, Max, Team, and seat-based Enterprise plans. The other half of the same announcement is the one that hits current usage: those plans already have a temporary 50% increase, and that promo only runs until the new baseline lands.

ClaudeDevs put it this way on August 29: “Starting September 14, we're permanently raising standard weekly limits in Claude Code by 25% for Pro, Max, Team, and seat-based Enterprise plans. Until then, the current 50% increase will be in place.”

Read against the old baseline, that is a lasting bump. Read against this week’s quota, it is a cut.

150 down to 125

BleepingComputer spelled out the ratio. If Claude Code’s original weekly allowance was 100, the temporary boost gives you 150 today. On September 14, that drops to 125.

You will have 25% more Claude Code usage than before the temporary promotion, and about 17% less than you have right now.

Anthropic later made the second number official. BleepingComputer reports that the company deleted the original thread and posted a clarification that admits the reduction: “Compared to today, this works out to a 17% reduction in weekly limits on Claude Code.”

The current 50% temporary increase remains available through September 13. The new permanent limits take effect on September 14.

That calendar is the planning fact. Teams that sized evals, refactors, or agent loops around the promotional weekly cap should treat September 14 as a real capacity change — not a rounding error, and not another stay of execution.

Who is in the 25% bucket

The ClaudeDevs post names four plan families: Pro, Max, Team, and seat-based Enterprise. Those are the plans that get the permanent 25% raise, and therefore the 17% step-down from today’s promo.

The posts do not describe other plan types, dollar prices, or a prompt-by-prompt quota. Anthropic also did not publish a single universal “you get N Claude Code messages” number. Usage, the company says, can vary based on conversation length, model choice, tool usage, and effort level, so the weekly allowance does not translate into a fixed number of Claude Code prompts.

Practically, two teams on the same plan can hit the weekly wall at very different ticket counts. Long threads, heavier models, more tool calls, and higher effort settings all spend the allowance faster. Shorter threads, lighter models, fewer tools, and lower effort stretch it. None of that changes the September 14 ratio. It only changes how quickly you burn whatever cap you have.

The visibility tease

Anthropic did not leave the clarification at 17%. In the same follow-up BleepingComputer quoted, the company said: “We’re working on exciting changes that will make it feel like you’re getting more from Claude, while having more visibility and control of your usage. Can’t wait to share them.”

Those changes are unspecified. No ship date, no feature names, and no claim that they restore the promotional weekly cap. Until they exist in product, they are not a substitute for the 17% you lose versus current limits.

If the follow-up is real instrumentation — seeing burn as it happens, steering which seats or jobs consume the week — that would help teams live inside a tighter cap. It would not, on the numbers Anthropic already published, put 150 back to 150.

Budget for the drop, not the headline

The first post led with a permanent 25% raise. That number is true versus the pre-promo standard. The number that matches how people are working under the current promo is the 17% reduction versus today.

Use the 100 → 150 → 125 sketch as a scaling template. You do not need Anthropic’s private baseline units. Take current weekly Claude Code headroom as 150 parts. After September 13, plan on 125 parts.

If you are already brushing the promotional ceiling, September 14 is when overflow starts earlier in the week. If you sit well under the promo cap, you may not feel it — unless a busy week would have fit at 150 and does not fit at 125.

Re-run that comparison on your own traces before the date. Look at conversation length, model mix, tool usage, and effort level, because those are the factors Anthropic says actually move consumption. Then watch ClaudeDevs for the visibility and control follow-up. The dated change is already on the calendar.

Read the August 29 ClaudeDevs post for the plan list and dates, then BleepingComputer’s recap for the 150-to-125 example and the 17% clarification — and size next month’s Claude Code budget off 125, not 150.