Anthropic admits Claude Code users hitting usage limits 'way faster than expected'
- Reference: 1774957511
- News link: https://www.theregister.co.uk/2026/03/31/anthropic_claude_code_limits/
- Source link:
Anthropic has acknowledged the issue, [1]stating that "people are hitting usage limits in Claude Code way faster than expected. We're actively investigating... it's the top priority for the team."
A user on the Claude Pro subscription ($200 annually) said on the company's Discord forum that "it's maxed out every Monday and resets at Saturday and it's been like that for a couple of weeks... out of 30 days I get to use Claude 12."
[2]
The Anthropic forum on Reddit is buzzing with complaints. "I used up Max 5 in 1 hour of working, before I could work 8 hours," [3]said one developer today. The Max 5 plan costs $100 per month.
[4]
[5]
There are several possible factors in the change. Last week, Anthropic [6]said it was reducing quotas during peak hours, a change that engineer Thariq Shihipar said would affect around 7 percent of users, while also claiming that "we've landed a lot of efficiency wins to offset this."
March 28 was also the last day of a Claude [7]promotion that doubled usage limits outside a six-hour peak window.
[8]
A third factor is that Claude Code may have bugs that increase token usage. A user [9]claimed that after reverse engineering the Claude Code binary, they "found two independent bugs that cause prompt cache to break, silently inflating costs by 10-20x." Some users confirmed that downgrading to an older version helped. "Downgrading to 2.1.34 made a very noticeable difference," [10]said one.
The [11]documentation on prompt caching says that the cache "significantly reduces processing time and costs for repetitive tasks or prompts with consistent elements." That said, the cache has only a five-minute lifetime, which means stopping for a short break, or not using Claude Code for a few minutes, results in higher costs on resumption.
[12]Contracts are in C++26 despite disagreement over their value
[13]Linear moves sideways to agentic AI as CEO declares issue tracking dead
[14]JetBrains shifts to agentic dev with Central, retires pair programming
[15]Mozilla introduces cq, describing it as 'Stack Overflow for agents'
Developers can upgrade the cache lifetime to one hour but "1-hour cache write tokens are 2 times the base input tokens price," the documentation states. A cache read token is 0.1 times the base price, so this is a key area for optimization.
Anthropic does not state the exact usage limits for its plans. For example, the Pro plan promises only "at least five times the usage per session compared to our free service." The Standard Team plan promises "1.25x more usage per session than the Pro plan." This makes it hard for developers to know what their usage limits are, other than by examining their dashboard showing how much quota they have consumed.
Problems like this are not unusual. Earlier this month, users of Google Antigravity were [16]protesting about similar issues.
[17]
Bugs aside, what we are seeing is an implicit negotiation between users and providers over what is an acceptable pricing and usage model for AI development. Users want to control costs and providers need to make a profit. There is also a disconnect between vendor marketing that urges developers to insert AI into every process, including in some cases automated workflows, and a quota system that can cause AI tools to stop responding.
"For folks running Claude Code in automated workflows: rate-limit errors need to be caught explicitly – they look like generic failures and will silently trigger retries. One session in a loop can drain your daily budget in minutes," [18]observed one user. ®
Get our [19]Tech Resources
[1] https://old.reddit.com/r/Anthropic/comments/1s7zfap/investigating_usage_limits_hitting_faster_than/
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/devops&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2acvvpFwVXtwoqCClhIOytAAAAIQ&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[3] https://old.reddit.com/r/Anthropic/comments/1s8ex9a/nothing_changed/odgoc24/
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/devops&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44acvvpFwVXtwoqCClhIOytAAAAIQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/devops&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33acvvpFwVXtwoqCClhIOytAAAAIQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[6] https://www.theregister.com/2026/03/26/anthropic_tweaks_usage_limits/
[7] https://support.claude.com/en/articles/14063676-claude-march-2026-usage-promotion
[8] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/devops&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44acvvpFwVXtwoqCClhIOytAAAAIQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[9] https://www.reddit.com/r/ClaudeCode/comments/1s7mitf/psa_claude_code_has_two_cache_bugs_that_can/
[10] https://www.reddit.com/r/ClaudeCode/comments/1s7mitf/comment/odbc868/
[11] https://platform.claude.com/docs/en/build-with-claude/prompt-caching
[12] https://www.theregister.com/2026/03/31/cplusplus26_approved/
[13] https://www.theregister.com/2026/03/26/linear_agent/
[14] https://www.theregister.com/2026/03/25/jetbrains_central/
[15] https://www.theregister.com/2026/03/24/mozilla_introduces_cq_stack_overflow/
[16] https://www.theregister.com/2026/03/12/users_protest_as_google_antigravity/
[17] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/devops&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33acvvpFwVXtwoqCClhIOytAAAAIQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[18] https://old.reddit.com/r/Anthropic/comments/1s7zfap/investigating_usage_limits_hitting_faster_than/odf4ipy/
[19] https://whitepapers.theregister.com/
Someone else also hoarded all ram so you can't have your own computer, but you must pay someone else to use theirs.
Funny how tokens always get used up quicker...
Wonder how often the bugs goes the other way, and user get to use way more then they are paying for?
Re: Funny how tokens always get used up quicker...
The token economy gives no incentive for Claude to generate quality output. Instead of oneshotting a proper solution, it generates crap ones so that users spend more.
The AI operators still lose money on each token....
I think it's safe to assume that we'll see all the differential pricing strategies that the mobile operators tried over the last 20 years (usage caps, time caps, priority pricing, etc.) repeated by the AI operators in fast forward mode in the coming months.
Until everyone realizes that only a $2500/month or upward subscription could sustain the big, centralized models if they are supposed to stop losing money, or even more if they should have any chance of paying interest on their past investments to their shareholders.
Re: The AI operators still lose money on each token....
As little as that?
Re: The AI operators still lose money on each token....
It will be interesting how high they push it. I'd expect it could hit 20K/mo for 8 hrs of day usage unlimited for those hours. And much like people, only so many tokens are going to get spit out per minute.
I'm sure they will raise prices slowly to fully boil the frog.
Re: The AI operators still lose money on each token....
They would if they weren't facing competition from China which has significantly lower overall costs.
Re: The AI operators still lose money on each token....
Assuming the initial start up cost to get the AI accelerators / GPUs / servers / racks / buildings / etc are all the same, which is the biggest ongoing running cost?
Is it man power? power? something else?
Depending on what it is, we can probably figure out which place has the lowest cost to run such DCs.
Re: The AI operators still lose money on each token....
They will need to check where the commercial optimum is, as every $ more costs them customers.
100k customers for $250/month result in $25M/month revenue
2k customers for $2500/month only result in $5M/month revenue, while saving some electricity and DC floor costs.
However, running AI services is probably not becoming the license to print money, as which it was advertised to investors...
Re: The AI operators still lose money on each token....
I suspect it is going to actually work out cheaper to employ a human than to pay the real cost of these so-called AI systems.
Re: The AI operators still lose money on each token....
@katrinab I think it'll go full circle.
For the short to medium term lots of humans will find what they did being replaced by AI. Once it's whipped up an inevitable shit storm humans will be brought back in to sort the mess out.
We've no way of knowing. It just seems grimly predictable.
Re: The AI operators still lose money on each token....
Yes, that's certainly how it feels. And when everyone cries "why did no-one tell us", all of us will do the classic face-palm gesture and say "BUT WE DID! You wouldn't listen!"
Colour me surprised...
... NOT.
This entire article could just as easily be describing Slot Machines; make of that what you will...
taxi analogy
It’s like taking a taxi with a driver who has no sense of direction and chooses the longest route, leading to unnecessary overspend.
Also the source code has just been released by error so others might be able to FTFY
Re: taxi analogy
..and the taxi driver has never been taught to drive - they've just been forced to watch every film and tv show that's ever been made that features vehicles
"breaks automated workflows"
Asking a bingo machine what to do next is some new definition of automation that I haven't heard of before.
Oh dear, it's somebody else's computer with all that entails.