Anthropic's ClaudeDevs account said Thursday that all users had received a rate-limit reset, the third blanket action reported in three weeks. The company cleared the five-hour and weekly counters without saying why. Max accounts keep assigned weekly reset times, while Pro and Max users reported different windows before their usual weekly refresh.
What Changed
- Anthropic reset Claude five-hour and weekly limits on July 16, the third blanket reset in three weeks.
- Fixed weekly schedules left some subscribers hours to use the refill while others had days.
- One user estimated a Friday reset could recover 86% of a weekly allowance while a Thursday reset recovered almost none.
- Claude Code weekly limits remain 50% higher and included Fable 5 access runs through July 19.
AI-generated summary, reviewed by an editor. More on our AI guidelines.
ClaudeDevs repeats its July 9 notice
On July 16, ClaudeDevs posted one sentence: "We've reset 5-hour and weekly rate limits for all users." A July 1 post linked that reset to Fable 5's return. ClaudeDevs then published the same unexplained statement on July 9; the July 9 and July 16 notices used identical wording.
Anthropic has explained earlier limit changes in several ways. In March, it tied peak-hour adjustments to demand and said about 7% of users would encounter session limits they had not reached before. Those changes applied from 5 a.m. to 11 a.m. Pacific on weekdays and left weekly limits unchanged. In May, the company linked higher limits to added computing capacity. Some earlier resets followed acknowledged limit bugs. Startup Fortune linked one July reset to a week of service incidents. Anthropic's July 16 announcement identified no cause.
After the July 9 reset, OpenAI's head of ChatGPT and Codex, Tibo Sottiaux, replied, "I smell fear." The reply came after OpenAI's GPT-5.6 Sol launch and offered a competitor's interpretation of the timing. Anthropic has disclosed no connection between the resets and OpenAI's releases.
Max accounts keep their assigned weekly clocks
Anthropic's current Max plan help page, updated July 13, says the weekly reset day and time are fixed for each account. The schedule stays the same regardless of when a customer starts using Claude or begins a subscription, and the account receives a full weekly allowance each cycle. The page also lists two weekly limits for Max accounts, one across all models and another for Sonnet. Reddit user diagrammatiks reported that the July blanket reset cleared the account's meter without moving its assigned deadline.
Reddit user stizzy6152, who had eight hours before a regular reset, wrote that the five-hour cap made using the full restoration impossible.
Get Implicator.ai in your inbox
Strategic AI news from San Francisco. No hype, no "AI will change everything" throat clearing. Just what moved, who won, and why it matters. Daily at 6am PST.
No spam. Unsubscribe anytime.
Under an even-use assumption, Reddit user Sunny-vibes calculated that a subscriber with a Friday reset could recover about 86% of a weekly allowance from a Thursday blanket reset before receiving the normal Friday reset. The user's corresponding estimate for a Sunday account was 57%, while an account whose regular reset arrived on Thursday could receive almost none. Anthropic did not produce the figures, which depend on a customer's consumption pattern.
The reset restored immediate access for users who had reached both caps. One commenter, wfh-without-pants, wrote that the five-hour, weekly and Fable meters had all been exhausted when the new allowance arrived.
Know someone who'd find this useful? ✉️ Email it to a friend in one click, or they can subscribe free here.
Thursday and Sunday users report different outcomes
lemonlemons wrote that a scheduled reset three hours after the blanket reset made it "useless to me." FrankPleau, whose regular reset was Sunday, said the sequence would give him a fifth full Fable run and then a sixth after his usual reset.
Anthropic's resets arrive amid separate disputes over allowance sizes it does not publish and a proposed class action contesting the Max tier multipliers. Anthropic has not connected either dispute to the July 16 reset.
Claude Code promotion runs through July 19
Help Net Security reported that Claude Code's weekly limits are temporarily 50% higher for Pro, Max, Team and legacy seat-based Enterprise users until 11:59 p.m. Pacific time on July 19. The increase covers Claude Code in the CLI, IDE extensions, desktop app and web. Other Claude products keep their existing limits. The promotion leaves five-hour limits unchanged; the July 16 blanket reset cleared both counters. Included Fable 5 access is scheduled to end at the same time on July 19.
Frequently Asked Questions
What did Anthropic reset on July 16?
Anthropic said it reset both the five-hour and weekly rate limits for all Claude users. The announcement did not say that plan prices, allowance formulas or assigned weekly reset schedules had changed.
Why did Anthropic reset the limits?
The July 16 notice gave no reason. Anthropic has tied earlier limit changes to demand, added computing capacity and software bugs, but it has not connected any of those explanations to this reset.
Why did the reset benefit subscribers differently?
Max accounts have assigned weekly reset times. Someone whose normal reset followed within hours had little time to spend the restored allowance, while another subscriber could use it for several days before the regular weekly refill.
How large could the difference be?
Under an even-use assumption, Reddit user Sunny-vibes estimated that a Friday-reset account could recover about 86% of a weekly allowance before its regular refill. A Thursday-reset account could recover almost none. Anthropic did not produce those figures.
What changes on July 19?
Claude Code weekly limits are scheduled to return from a temporary 50% increase after 11:59 p.m. Pacific time on July 19. Included Fable 5 access is scheduled to end at the same time; the promotion itself does not change five-hour limits.
AI-generated summary, reviewed by an editor. More on our AI guidelines.



IMPLICATOR