Anthropic Announced today that effective September 14, there will be a permanent increase. Claude Code Weekly usage quota of 25%, applicable to Pro, Max, Team, and seat-based Enterprise plans. The tweet received over 16,000 likes on X, but developers quickly calculated what the official announcement didn’t make clear in the main post: compared to the temporary quota users actually currently have, this is a 17% reduction. Users on X even attached a Community Note to the main tweet, directly labeling this as a downgrade rather than an upgrade.
How did the 17% shrinkage figure come about?
The story starts in May this year. On May 13, Anthropic introduced a 50% increase in Claude Code’s weekly quota.Limited-time offerOriginally planned for one month. But the promotion was extended four times in total, ultimately pushed to August 31, letting users gradually get used to the 150% quota level. During this period, Anthropic also reached a compute cooperation agreement with SpaceX, simultaneously doubling Claude Code’s five-hour rolling limit and removing peak-hour quota reductions for the Pro and Max plans.
Now, Anthropic announced“Permanently increase by 25%,” starting from the original baseline (100%).calculation, rather than based on the temporary quota (150%) the user currently has. The conversion is as follows:
- original standard limit：100%
- Current temporary promotional quota: 150% (original baseline +50%)
- Starting from September 14, permanent quota: 125% (original baseline +25%)
- Compared to the current: shrink by about 17%
Anthropic emphasized “permanently increased by 25%” in the main post, but hid the key detail that it was actually a 17% cut compared to current levels in the reply thread. Users quickly attached an X Community Note to the main post, warning others that this was effectively a downgrade.
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.
— ClaudeDevs (@ClaudeDevs) August 29, 2026
Behind the Promotions: Stopgap Measures for the April Quality Crisis
This 50% promotion was not introduced without reason. Between March and April 2026, Claude Code faced a serious quality crisis. Anthropic later admitted that three problems occurred consecutively during this period: on March 4, the default reasoning intensity was lowered from high to medium, resulting in a drop in output quality; on March 26, a bug caused sessions idle for more than an hour to continuously have their thinking content cleared, making the model appear “forgetful” and repetitive; on April 16, a system prompt added to reduce verbosity actually harmed code quality.
Because the three issues affected different traffic scopes and timeframes, the overall effect looked like a broad quality degradation. Making matters worse, Anthropic’s internal usage data and evaluation benchmarks initially couldn’t reproduce the problems users reported, which dragged out the root-cause identification process for nearly two months. On April 23, Anthropic published a post-mortem acknowledging that this wasn’t the experience users deserved, and reset all subscribers’ usage quotas as compensation. The 50% promotion in May was exactly to win back users who had churned due to the quality issues.
Competitors are also tightening: OpenAI has simultaneously reinstated strict restrictions.
Anthropic isn’t the only company adjusting its usage limits. Just a few days ago, OpenAI also reinstated a strict five-hour usage limit for Codex and ChatGPT Work plans, claiming the same quotas should now stretch 10% to 50% further. Both companies tightening limits at the same time signals that the “unlimited usage” fantasy of subscription-based AI tools is fading, as players begin confronting the reality of compute costs and sustainability. For developers, this means the overage-friendly experience they’ve enjoyed over the past few months is coming to an end, and they’ll need to reassess the value for money across platforms.
Anthropic’s next step
Anthropic said the adjusted quota is a level that has been assessed as sustainable to offer. To make up for the 17% reduction, the company previewed that it is developing new interface features to give users better visibility and control over their remaining quota, while also hinting at improvements that will make users “feel they get more,” though it has not yet revealed specific details.
Additionally, this adjustment only affects Claude Code (CLI, IDE extension, desktop, and web versions), and usage allowances for the Claude chat interface are not affected. Furthermore, starting September 14, the separate weekly limit for the Fable model will be removed, allowing users to freely allocate their full allowance across their preferred Claude models.
From the damage-control promotion in May to the new permanent policy in September, Anthropic’s quota strategy over these four months has followed a path of “raise first, then take back.” The official rhetoric is a “permanent increase,” but for developers who rely on Claude Code every day, the quota available after September 14 has indeed decreased. At a time when competitors like Cursor, Windsurf, and GitHub Copilot are all expanding quotas, this 17% gap is enough to make some heavy users reevaluate their options. Whether Anthropic can close this gap with its upcoming usage management tools will determine whether this adjustment is a “smart sustainable strategy” or “the beginning of user churn.”
Source: KOCPC Chinese