Cases
Newest first. Official receipts and pattern files in one list.
Spend limit reached, except it was not
The accountant in the stack invented a ceiling and then enforced it.
On 7 Oct 2026 status.claude.com said some organizations were incorrectly paused for reaching a spend limit. Requests were refused on the Claude API, Claude.ai, Claude Code, Claude Security, and Claude Cowork. Marked resolved the same day at 21:23 UTC.
The usage page could not load the usage
The meter broke. The work did not. You just could not see the bill accruing.
From 00:00 to 01:40 UTC on 7 Oct 2026 many requests to load usage data on platform.claude.com and the Admin API usage report returned errors. Messages API requests were not affected. A later same-day incident left recent usage delayed by about 30 minutes while a full fix was still in progress.
Start a deck and the next two weeks cost half
The discount lives inside the conversation that began as a slide. Leave the thread and you are on full fare.
On 1 Oct 2026 @claudeai said that for two weeks, starting a design, deck, or doc in the Claude app makes the work that follows in that conversation use 50% less of usage limits. The post points people at Claude Sonnet 5.5.
October rerun: bought credits, balance still late
September had this bug. October kept the template.
On 1 Oct 2026 status.claude.com investigated delays applying purchased credits to account balances. Some API requests failed for insufficient credit. A fix was in monitoring by 18:23 UTC and marked resolved at 22:36 UTC. Console and API were affected.
Do not sign out: an hour when login was the outage
The official advice was stay where you are. Logging in had become a feature you might not get back.
On 29 Sep 2026 elevated errors hit claude.ai (web, desktop, mobile), Claude Code, Claude Cowork, the console, and the API from 14:00 to 14:59 UTC. A second issue blocked sign-in, including SSO and Sign in with Apple, plus new chats, voice, purchases, and uploads. Some messages sent in that hour may not have been saved.
Hit the 5-hour wall and Claude Code tries to land the plane
The cutoff still happens. It just borrows next week's meter to finish the sentence.
On 25 Sep 2026 @ClaudeDevs said Claude Code will try to find a graceful stopping point when a 5-hour limit hits mid-task, instead of cutting off mid-edit. The wrap-up uses a small, fixed allowance pulled from the weekly limit.
Opus 5.5: 20% more session, 25% further, one reset in the drawer
A bigger glass, a cheaper pour, and a spare refill you are trusted to open yourself.
On 22 Sep 2026 Anthropic launched Claude Opus 5.5 and raised five-hour usage limits on Pro, Max, and Team. @ClaudeDevs said session limits increase 20% that day, Opus 5.5 goes 25% further inside limits because it is priced lower than Opus 5, and paid plans get a reset they can apply any time until 22 Oct.
You bought credits. The API still said you were broke.
The card cleared. The balance did not. The model sent a poverty notice.
From 12:10 to 21:35 UTC on 1 Sep 2026, users who hit a zero usage-credit balance saw delays before newly purchased credits became available. Some Claude API requests returned a credit-balance-too-low error in error. Status marked it resolved on 2 Sep.
Fable 5.1 ships, and the meters go back to zero
Launch day and refill day, same sentence. The product note is also the apology.
On 1 Sep 2026 @ClaudeDevs said that with Fable 5.1 out that day, 5-hour and weekly limits had been reset for all users.
Mythos, Fable, Opus: a three-hour error window
They named the models. They did not name the cause. The menu was the postmortem.
Status page reported elevated errors on Claude Mythos 5, Fable 5, Opus 5, Opus 4.8, and other models from 04:50 to 07:36 UTC on 24 Aug 2026. Cause identified, then resolved. Surfaces included claude.ai, the API, Claude Code, and Claude Cowork.
Opus 5 and Haiku 4.5, identified, fixed, next slide
Cause identified. Cause unnamed. Users already know the error code.
Status page identified a cause of elevated errors on Claude Opus 5 and Claude Haiku 4.5, implemented a fix, and marked the incident resolved the same morning UTC. Affected claude.ai, API, Claude Code, and Claude Cowork. No public root-cause detail beyond that sequence.
Weekly Code limits stay 50% higher, now through a newer month
Permanent is the destination. August 31 is the layover they keep booking.
@ClaudeDevs extended the 50% increase to weekly Claude Code limits through August 31 for the same plan families previously covered. The post hopes to make the change permanent, and notes capacity may stay tight from strong demand.
The family portrait of elevated errors, then just Opus
Start with the whole catalog. End with one model and a tidy UTC range.
Status page investigated degraded performance across multiple models, listing Mythos 5, Fable 5, Opus 5, Sonnet 5, Haiku 4.5, and others. Later updates narrowed to Opus 5. Official impact window 16:11 to 18:23 UTC. Resolved same day.
Opus recovered. Sonnet stayed for the second act.
Two models, one incident, staggered dignity.
Status page investigated elevated errors on Claude Opus 5, then reported Opus operational while Claude Sonnet 5 stayed degraded. Fix implemented and incident resolved the same afternoon UTC. Affected claude.ai, API, Claude Code, and Claude Cowork.
Login first, then the rest of the product remembered it existed
If you cannot sign in, every other feature is a rumor.
Status page investigated an authentication issue for claude.ai, Claude Code, and Claude Cowork, then widened to degraded performance on those surfaces plus platform.claude.com and the Claude API. Fix deployed and marked resolved the same evening UTC.
The safety classifier finally stopped billing you for the lecture
A small number of extra tokens, until someone notices the meter.
@ClaudeDevs said Auto mode runs a classifier on every tool call and that as of 7 Aug 2026 that classifier overhead no longer counts toward usage limits on Pro, Max, and Team. Enterprise and API get the change once Auto mode is default for those users.
Fable 5 biology safeguards: 85% fewer 'I cannot help with that'
Everyday health questions get a wider hallway. Virology still hits the velvet rope.
@claudeai said it updated Claude Fable 5 biology safeguards to reduce false positives. Internal testing claimed about 85% fewer biology-related fallbacks across product surfaces. Dual-use topics still fall back to Opus 5; Fable is still not for professional biology research and drug development.
Two network failures, one status page, zero limit reset
Network karaoke: 'stable right now' is the new encore.
@ClaudeDevs reported two incidents in 24 hours with elevated errors or reduced availability. Multiple separate network failures cut capacity; some requests failed during traffic reroute. Capacity partially restored, not fully clear. Status page logged multi-hour elevated errors on Jul 29 and again on Jul 30 across many models.
Documents, Code, Cowork, Tag, Design: a feature buffet of downtime
When half the product surface shares one bad afternoon, the status page becomes a roll call.
Status page reported a service disruption affecting document creation on claude.ai, Cowork Remote, Claude Code, Claude Code on the Web, Claude Tag, and Claude Design. Identified and resolved the same afternoon UTC.
Max plan, meet usage credits you already paid not to need
Included in plan, until the meter disagrees.
Status page: some Claude Code Max plan users were incorrectly prompted to use usage credits for Fable 5, which is included in their plan. Fix landed; restart Claude Code if affected. Same day Fable rolled into Max and Team Premium at 50 percent limits.
Fable demand was 'challenging to predict' (everyone predicted it)
Capacity theater, act three: the encore is a credit and a percentage.
Fable 5 rolled into Max and Team Premium at 50 percent of limits from July 20. Pro and Team Standard get usage credits and a one-time $100 credit. Staged access blamed on unpredictable demand. Claude Code weekly limits stay 50 percent higher through August 19.
Fable goes missing for 30 minutes, wallet has feelings
The model left the dropdown. Some invoices did not get the memo.
Fable was not selectable on claude.ai or Claude Code for about 30 minutes. Users may need to restart Claude Code and reselect with /model. Separate note: if extra usage was on, possible overcharge; refunds plus matching additional credit promised.
We've reset 5-hour and weekly rate limits for all users (again)
The shortest incident report format in the industry: one sentence and a full meter.
Brief official post resetting 5-hour and weekly rate limits for all users, without a detailed public root cause in the tweet body.
The six-section essay that refuses to decide
It depends, said the model that was hired to decide.
Asked for a sharp take on a narrow product decision. Received balanced mush that could sell any product.
Fable access and Code weekly limits get another calendar extension
Demand still challenging. The extension team is not.
Extension of Claude Fable 5 access on all paid plans and Claude Code weekly rate limits 50% higher through July 19. Same 50% weekly Fable usage framing as prior promo posts.
Code that demos the happy path and lies on the edges
It type-checks. Your production pager still has a career.
Asked for a small utility with validation and error handling. Happy path looked clean. Edges were vibes.
You're absolutely right, and then nothing useful happens
Agreement is free. Correctness still wants a receipt.
User pushed back on a technical claim with a short correction. Model flipped into praise mode without re-checking.
We've reset 5-hour and weekly rate limits for all users (Jul 9 edition)
Midweek reboot, no postmortem, full meter. The genre writes itself.
@ClaudeDevs posted a global reset of 5-hour and weekly rate limits for all users, with no root-cause detail in the tweet body.
Fable 5 access extended again, calendar moves to July 12
Another extension. The spreadsheet of end dates is getting longer than the model card.
@claudeai extended access to Claude Fable 5 on all paid plans through July 12, with the familiar 50% weekly usage cap then credits or other models.
Fable 5 is back, with new safeguards and a 50% weekly cap
Welcome back. Half your week of the shiny model. The other half is Opus when a classifier gets nervous.
claudeai announced Fable 5 is back. Thread covers updated cybersecurity safeguards after US government conversations (more false positives short term, fallback to Opus 4.8), unchanged biology/chemistry classifiers still broader than desired, and promotional access on paid plans through July 7 at up to 50% of weekly usage then credits or other models.
About 3% saw a fake weekly cap; some got blocked for real
The UI invented a limit. The product enforced the fan fiction.
Bug hit roughly 3% of Claude Code Max and Pro users with an incorrect weekly usage limit display, sometimes blocking messages. Fixed with limit resets for affected users; follow-up reset expanded to everyone across all plans.
Another weekend, another 'we've reset limits for all users'
The most reliable Claude Code ritual of summer 2026 was not a slash command.
Short official post resetting 5-hour and weekly rate limits for all users, without a long incident narrative in the tweet body.
Enjoy Fable 5: a global limit reset for the launch afterparty
New model, clean meter, same recurring genre of tweet.
@ClaudeDevs reset 5-hour and weekly rate limits for all users with a note to enjoy Fable 5, during the Fable launch window.
Subagents threw a house party on your weekly quota
The agents were productive. Your remaining limit was not invited.
Fix for Claude Code sessions spawning excessive parallel subagents and burning usage faster than expected. 5-hour and weekly rate limits reset for all Pro and Max users.
Claude Code responsiveness tour, conducted by the same team that felt the heat
A progress report after the fire is still a progress report. We clap and we archive.
Broad update on efforts to make Claude Code more responsive and reliable, summarizing work done.
Happy Friday: everyone's limits just got a free reboot
Incident response, weekend edition: skip the postmortem, ship the reset.
@ClaudeDevs reset everyone's 5-hour and weekly rate limits with a casual Friday note and no detailed root-cause story in the post.
Claude Code weekly limits up 50% through mid-July
Promotional generosity with an expiration date. The meter still has a birthday.
Claude Code weekly limits increased 50% through 13 Jul 2026 for Pro, Max, Team, and seat-based Enterprise. Applies CLI, IDE, desktop, web. No opt-in. Stacks with prior 2x five-hour limit increase.
50+ reliability fixes last week, 60+ this week, still counting
Nothing says 'we heard you' like shipping fixes by the dozen like confetti.
Wave of Claude Code reliability fixes covering long sessions, agent loop efficiency, auth environments, and terminal issues.
Limits go up, peak-hour haircut goes away, API Opus breathes
Nothing says 'we heard the queue' like shipping more tokens and a partnership press line.
Official raise of Claude Code 5-hour limits (2x) for Pro/Max/Team/seat Enterprise, removal of peak-hours limit reduction on Code for Pro/Max, and higher API rate limits for Opus. Related Anthropic note on higher limits and SpaceX compute.
Claude Code harness drama, models plead not guilty
The brain was fine. The body that types into your repo had opinions.
After users reported Claude Code quality had slipped, @ClaudeDevs published a postmortem on three issues in the Claude Code / Agent SDK harness (also impacting Cowork). Fixed in v2.1.116+. Models and Claude API said not regressed. Usage limits reset for subscribers.
Opus 4.7 called normal code edits malware (stale safety prompt)
Not cautious. Just haunted by last week's system prompt.
@ClaudeDevs admitted some users hit Opus 4.7 refusing normal code edits with malware warnings. Bug on their side: older builds applied a stale safety prompt Opus 4.7 does not need. Fix: claude update or relaunch.
Long context met short math: Opus rate limits get a hotfix and a reset
The model got a bigger window. The meter kept using the old ruler.
Official fix for subscription rate limits that were not properly adjusted for long context requests on Opus 4.7, plus a reset of 5-hour and weekly limits.
Thank-you double usage outside peak hours, two weeks only
Gratitude with an expiration date and a peak-hours asterisk.
@claudeai announced doubling usage outside peak hours for the next two weeks as a thank you to Claude users.
Cowork does more complex work; Pro may hit limits sooner
Feature intensity is free. The meter still bills by the sip.
While shipping Cowork research-preview improvements, @claudeai noted that because Cowork handles more complex work, Pro users may hit their usage limit sooner.
Max 5x holiday promo glitched; fixed with a limit reset for the affected
The gift receipt had a bug. They fixed the promo. They reset the meter. Merry glitchmas.
@claudeai identified and fixed an issue with the usage promotion for Max 5x users, and reset usage limits for all affected accounts.
Holiday double limits through New Year's Eve, gift-wrapped capacity
Season's greetings from the meter. Enjoy the temporary abundance.
Starting midnight PT Dec 24, all Pro and Max plans received 2x usual usage limits through New Year's Eve.
Three infrastructure bugs get a full postmortem era
The rare AI post where 'it was the plumbing' is both true and not a personality excuse.
Detailed postmortem covering three infrastructure bugs that hit Claude between August and early September 2025.
Two quality bugs, one grateful nod to the bug report factory
Translation: the model was not vibing. Community screenshots did the QA department's job.
Official note that two issues had been affecting quality in some Claude responses, with monitoring still ongoing.