Claude officially announced a permanent 25% credit limit increase, but the actual amount received was 17% less?
Anthropic announced a permanent 25% increase in the standard weekly credit limit for Claude Code starting September 14th, but the previous temporary 50% increase will end, resulting in an actual decrease of approximately 17% in available credit. On the same day, OpenAI fixed a bug related to credit limit consumption in the Codex backend and reset credit limits.
Key Facts
According to QuantumBit, Anthropic announced on September 1st that it will begin implementing [measures] starting September 14th.Permanently increase Claude Code standard weekly credit limit by 25%.It covers paid tiers such as Pro, Max, Team, and Enterprise. However, Claude Code is currently implementing a round of...Temporary 50% increase in quotaThis bonus will end on September 14th, at which time it will be replaced by a permanent +25% bonus. If the base credit limit is 100, the current usable limit is 150, and after the adjustment, it will be 125.The actual amount received was reduced by approximately 17%.Anthropic officials acknowledged the change in the comments section, stating that they are optimizing it to improve perceived value and usage visibility for users.
Background and Impact
The announcement sparked user dissatisfaction, with numerous criticisms appearing in the comments section. Some users indicated they were considering unsubscribing, while others compared Claude to Chinese models, accusing it of "tightening its grip against the trend." Meanwhile, several hours later, Tibo, head of OpenAI's Codex, announced a reset of Codex quotas and disclosed that several bugs related to background quota consumption had been fixed, including context compression, memory process infinite loops, continued execution after task completion, abnormal frequency of automatic tasks, sub-agents calling high-cost models, and repeated historical summaries. He predicted that the available quota would increase by 10% to 50% after the fixes. In contrast, public opinion believes Anthropic lost points in its communication strategy. Some analysts suggest that Claude's move may be related to computing power costs and user structure, but there is no publicly available data to support this claim.
Restrictions and Sources
This article is based on a report by Qbitai. Specific quota calculations and user feedback are cited from the original article and have not been independently verified. Details of Anthropic and OpenAI's official announcements, fixes, and user data are subject to the current information on their official websites. The Grok assessments and developer feedback mentioned in the article are isolated cases and do not constitute general conclusions. Source: Qbitai (www.qbitai.com), Link:https://www.qbitai.com/2026/09/482406.html。