This does not explain the changes to documentation.
> When we do land on something, if it affects existing subscribers you'll get plenty of notice before anything changes. Will hear it from us, not a screenshot on X or Reddit.
If you don't want things like this spreading through screenshots of X and Reddit, don't run "tests" like this in the first place!
(Also "if it affects existing subscribers" is a cop-out, I need to know the pricing of Claude Code for NEW subscribers if I'm going to adopt it at a company with a growing team, or recommend it to other people, write tutorials etc.)
I can't trust Anthropic to manage their products in a way that supports my workflow.
ive been trying to make the case all year that if we're going to let employees do shit with ai, lets try claude. in the past like.. 2-3 weeks all that goodwill has basically evaporated.
local inference needs to take off asap because all of these entities actually suck and i wouldn't trust a single sla with anthropic. they are not acting like a serious company right now, this is a joke.
No serious business uses Pro or Max, they are all on Anthropic API billing.
In fact with this move it is plainly obvious that Anthropic is moving compute from prosumers towards enterprise.
It will be interesting to see how it all plays out, but I suspect if cost continues to increase and output only improves incrementally from here, that the cost will be the final decider rather than the competence.
I could see it being a thing we use only sometimes, for some things, but ultimately remain reliant on developers to get the work through the pipeline.
Larger companies are using Claude through AWS Bedrock and are willing to easily pay $5k+ per engineer per month for it.
Developer salaries are driven up by scarcity - scarcity of developer skills overall and scarcity of developer skills in specific places like California. If AI models destroy the scarcity then the price worth paying for a coding agent will drop dramatically.
Maybe Anthropic can get away with it for a couple of months. But this will not last.
This not nothing.
With Sonnet it's a bit better, but I can get the same performance with GPT-5.4.
Now I'm pretty much paying the 20€ for Claude Pro so it can plan/review stuff and then I use pi.dev + GPT-5.4 for the actual work.
That said, I seem to be caught in that 2% test if I open in a private tab. What nonsense. I wouldn't be paying for Claude if it wasn't for its quality abilities, which necessarily includes Claude Code.
I find that with Opus 4.7 I can do two messages. Once I had a short session with 4-5 messages and it consumed $10 in extra usage.
This relegated Claude to a backup option in addition to Codex, which has the better desktop app anyway, and much better usage limits.
I’m considering to even cancel Claude entirely.
I agree, but can you really use Claude Code on the Pro plan as a full time developer, or professional 'knowledge worker' without hitting the usage limits fairly early in the day anyway?
I'm in the academia, and Claude's performance in my field could be described as a very fast junior grad student. When I use Claude Code, I typically spend a few hours figuring out what needs to be done exactly, and describing it in sufficient detail. Then Claude does it in 30 minutes, while an actual student would need days. And then I spend anything from minutes to days evaluating the results, depending on if it needs to be tested with real data and how much weirdness those tests uncover.
But I also have other work to do beyond guiding the automated grad student. Which means my Claude Code usage rarely exceeds 1–2 hours/week.
I have Pro Claude, Plus GPT and Pro Gemini. When one runs out I switch to another project on the next LLM. If I really need a task finished I'll restart it on another LLM, but I'm loathe to do that as it eats tokens just getting back up to speed.
A/B testing people without their informed consent is immoral, unethical, and should be illegal.
so, what i'm saying is : I think a lot of companies align themselves with the cash first and then measure whether or not the negative image/user impact is manageable .
(in fact I know they operate this way.)
Sure. Let me just A/B test whether or not you'll respond positively or negatively to having your news delivered via push notification or delayed by 10 minutes.
I'm sure you would appreciate being tested on without your consent, just so that I can make an extra quick buck at your expense. Nothing amoral or unethical about it.
It seems weird to segment this way though. Surely it’s better to just give Sonnet to your bottom tier, rather than cut out the entire Claide Code product entirely?
Give folks a taste rather than lock the whole product behind a $100/mo plan.
his title should be changed to Head of Corporate Bullshitting
They're hitting the physical limits of energy production and chip supply for inference capacity. There's literally nothing that can be done but reduce usage to spread it around for now.
And with no free trial period on top of that, nobody is going to want to pay $100+ just to check it out. I can't imagine the conversion rate of that test being positive.
I imagine Anthropic is trying to see how many users they can push to higher tiers with these new squeezes.
I hate to say it but I imagine it will work.
It’s going to suck for me, because I had gotten used to ridiculously cheap tokens, but I guess the era of subsidized tokens is over.
Until they go public, we are all just guessing.
It's hard to tell, honestly - about half the HN population will tell you that all the token providers are running inference at a profit when using the API and only the subscriptions are subsidised, while the other half will tell you that everything, including both the API and the subscriptions, are subsidised (i.e. running at a loss).
I work for a real business and switched from API billing to max+overflow. It saves money. It’s crazy not to. What are you talking about?
You may also have a very narrow view of how the world actually works, left as an exercise to the reader to figure out which one it is
I, and everyone else I have asked, see this new updated sales UI; sounds like more than 2%.
Hope you can still resume working on your projects without AI.
This is concerning though. If I lose my current usage allotment at this price point I will likely switch to codex
It also forces you to keep your workflow mostly harness-independent because Claude supports next to no standards and Codex does some.
Once they get people hooked, deskilled, and paying, the money ratchet only tightens.
And the companies KNOW that theyre replacing engineers, or trying to. So each engineer replaced is X salary a year they now have available, so make it back in SaaS LLM tokens.
I thought inference was cheap so there was little marginal cost of a new subscriber.
Based on how much money Zitron has reported that these companies are losing on every subscription, this feels more like they're just trying to survive. In other words "ohshittification."
Brilliant coinage, if it’s yours, congrats!
My take: it is not enshittification to raise the price for a product whose demand outstrips its supply. That is basic economics. There are alternatives, it’s not a monopoly. If you think it’s the best product, then pay more for it.
Personally I would be perfectly content if the price of Max went up a bit and Pro no longer worked for CC if it meant that Max was faster and more stable.
I had a bit of an epiphany the other day thinking about these VC companies offering products to the public at unsustainable prices. It's classic anticompetitive behavior.
You imagine anticompetitive behavior to come from a monopoly because they can afford to burn money to drive competition out before they bring prices back to profitable but the whole VC burn is the same thing. People talk about it a lot without really saying it explicitly when they talk about moats. The only moat Anthropic and OpenAI have is money and they utilize it by offering products below cost.
The two companies are just trying to outlast the other one until they are the only one left.
So it's not really enshitification as much as you were previously getting the deal of a lifetime.
There are some predatory pricing laws, but they're much more narrow than most people believe. There is no law requiring things to be sold for more than it costs to produce.
I think it's funny that these topics make people angry enough to demand that we make laws to force companies to raise prices. We'll stick it to these companies by forcing them to charge us more! That will show them!
Such laws would be very bad for startups and newcomers because they'd be forced to price their new product higher than established competitors who have economies of scale. It would be a nice handout to the big companies.
This is dumping and it is international trade. Maybe you don't realize it because you're American and have internalized it as business as usual.
These companies probably need to be forced to at least try to price their products at a level that would be sustainable long term.
I think it's funny that we're getting subsidized and discounted services and this makes some people so angry that the comment section is demanding laws that would force companies to charge us more.
> It's classic anticompetitive behavior.
Plenty of Pro subscribers never touch claude-code.
---
> For clarity, we're running a small test on ~2% of new prosumer signups. Existing Pro and Max subscribers aren't affected.
> When we launched Max a year ago, it didn't include Claude Code, Cowork didn't exist, and agents that run for hours weren't a thing. Max was designed for heavy chat usage, that's it.
> Since then, we bundled Claude Code into Max and it took off after Opus 4. Cowork landed. Long-running async agents are now everyday workflows. The way people actually use a Claude subscription has changed fundamentally.
> Engagement per subscriber is way up. We've made small adjustments along the way (weekly caps, tighter limits at peak), but usage has changed a lot and our current plans weren't built for this.
> So we're looking at different options to keep delivering a great experience for users. We don't know exactly what those look like yet - that's what we're testing and getting feedback on right now.
> When we do land on something, if it affects existing subscribers you'll get plenty of notice before anything changes. Will hear it from us, not a screenshot on X or Reddit.
https://x.com/TheAmolAvasare/status/2046724659039932830
---
Personally I love how they have increased everyone's quotas to counteract the Opus 4.7 tokenizer change a few days ago, but are immediately regretting it and trying to cut off subscription users.
If the subscriptions are unprofitable, then just communicate honestly, raise the price or lower limits for new subscribers transparently, and grandfather in existing users. That's what GLM coding plan is doing and it works fine for them. Don't ruin your reputation with opaque messaging and hidden changes. Lol
This. Why do so many companies fail to get this? Anthropic's user base, in particular, is intelligent enough to understand their constraints.
I think you may be overestimating the willingness of people to understand Anthropic’s concerns.
peak siliconbromaxxing
Like, they're just advertising different terms to test how many people would still click on it and very likely start the subscription process, but after they click they go back to the usual terms. Changing the whole payment flow, account models and permissions in their backend just for a quick test is usually too much work.
But yes, basically, if you're B and not A, and B has objectively worse terms than A, then you're just unlucky. But this is the essence of A/B tests. They are done by basically every company everytime, because it's the most straightforward and simple way to test new terms or designs.
After seeing my own issues with 4.6 and the mega-post on Github about declining metrics in a decent dataset of claude chats by Stella Laurenzo at AMD (https://github.com/anthropics/claude-code/issues/42796), I downgraded to the $100 plan. Hallucinations. Laziness. Lack of thinking. The responses on those mega-threads from Anthropic rubbed me the wrong way in a "you're holding it wrong" kinda way.
In the past week, I downgraded back to the $20 plan because the Codex $20 plan on 5.4 was working so well for me.
Then throw in other oddball events like the source code leak, and the super positive Anthropic events like their interactions with the current administration. It's a wild ride.
I can't understand removing Claude Code from $20. I'm interested to see whether this is confirmed or not.
I'm a career engineer and I went from being one of their most outspoken proponents (at least within my circle) and now.... I'm not.
Anthropic really pissed me off with their harness crap. They're well within their rights but their communication over it was enough to get me to swap. I don't need extra hurdles when there's a perfectly valid alternative right there. They don't have the advantage they think they do.
But if you’re stuck with no better model, it’s better than local models and no models.
I have to say, OpenCode’s OpenUI has taught me what modern TUIs can be like. Claude’s TUI feels more like it’s been grown than designed. I’m playing around with TUI widgets trying to recreate and improve that experience
Claude's TUI is not a TUI. It's the most WTF thing ever: the TUI is actually a GUI. A headless browser shipped the TUI that, in real-time, renders the entire screen, scrolls to the bottom, and converts that to text mode. There are several serious issues and I'll mention two that do utterly piss me off...
1. Insane "jumping" around where the text "scrolls back" then scrolls back down to your prompt: at this point, seen the crazy hack that TUI is, if you tell me the text jumping around in the TUI is because they're simulating mouse clicks on the scrollbar I would't be surprised. If I'm not mistaken we've seen people "fixing" this by patching other programs (tmux ?).
2. What you see in the TUI is not the output of the model. That is, to me, the most insane of it all. They're literally changing characters between their headlessly rendered GUI and the TUI.
> Claude’s TUI feels more like it’s been grown than designed.
"grown" or "hacked" are way too nice words for the monstrosity that Claude's TUI is.
Codex is described as a: "Lightweight coding agent that runs in your terminal". It's 95%+ Rust code. I wonder if the "lightweight" is a stab at the monstrosity that Claude's TUI is.
it goes into detail about llama-server args; quants to try; and layer/kv cache splits. I plan to try the techniques there.
Remember the old saying about boiling a frog? LLM corporations need to make most of their users pay hundreds per month, asap. This is Anthropic increasing temperature regulator under the pot just a tiny little bit. Not the first and not the last time.
> LLM corporations need to make most of their users pay hundreds per month, asap.
I think they need to at least have a 1 month introductory rate for the max plan at $20, or devs that decide to try out agentic coding just won't go to Anthropic.
That leads to downstream impacts, like when a company is deciding which AI coding tools to provide and the feedback management hears everyone is already used to (e.x.) Codex, then Anthropic starts losing the enterprise side of things.
I suspect a lot of people are like me. They got into this at the $20/month level individually to check things out. I'm not stressing things out, so I haven't moved up, but the moment I bump into a limit, I'll pull the trigger by default. Until then, I'm the sleeping dog, and you should let me lie.
Well, Anthropic decided to kick me. Now, I'm investing the time to figure out how to use the "open" and "Chinese" models assuming that Anthropic is about to screw me. Once I switch, Anthropic is going to have to demonstrate significant improvements over what I'm now using to get me to even consider them again.
Not according to their webpage: "Claude Code is included in your Pro plan. Perfect for short coding sprints in small codebases with access to both Sonnet 4.6 and Opus 4.7." [1]
> I'm a career engineer
I'm trying really hard here to be nice, but what the hell are you doing? Are you vibe coding multiple apps in parallel and calling it engineering?
Is it like those people who eat 2-3x the amount of meat to ensure they offset the positive impact 1-2 vegans are having? :D
Anthropic bleeds money per user. No matter if it's the $20 or $200 plan, every Claude Code user is unprofitable.
The only way to not bleed money is to eventually move everyone to API pricing. Hiring a personal senior engineer will likely be cheaper.
I'm not challenging your opinion, but this is an outlier in the general current public opinion about it.
EDIT: it is also surprising to me that everyone seems to believe the people at Anthropic are simply incompetent and recklessly risking their good reputation, while very few consider the possible good reasons they might have for taking such drastic measures. And I don't think it's because of financial pressures in their case
When I threw it at a difficult issue in an iOS app, it like GPT came up with wrongly guessed explanations. It only found the issue after I had it instrument the app and add extensive logs. Usually GPT 5.4 is the same.
Only that with GPT 5.4 it’s at least included in my subscription, while sending 3-4 messages to Opus 4.7 for this blew through my $20 plan limits and consumed $10 of extra usage on top. At that point I can’t help but bring up how much more expensive it is.
Rest assured OpenAI won’t want to leave that kind of money on the table…
With that much competition and ongoing improvements, I don’t have such a pessimistic view on future usage limits and cost.
At the speed everything is advancing I don’t think it’s such an advantage. They all catch each other up pretty fast. That’s why I prefer to pay Cursors and have access to all of them instead of being lock to a single one (even if that means to lose some discounted credits). If they opened Mythos today at a good price that would be something but that’s not the case and it won’t happen.
That's not how you keep your customers. None of these agents have a moat, I moved away from Cursor when they started doing what Anthropic is doing now, and never went back even when I was a paying customer since the start.
Yes, confirmed directly from Anthropic's website. Claude Code is included in the Pro plan, described as perfect for short coding sprints in small codebases with access to both Sonnet 4.6 and Opus 4.7. Claude
You said: Is claude code included with the pro subscription Is claude code included with the pro subscription
3:11 PM Claude responded: Yes, Claude Code is included with the Pro plan ($20/month). Yes, Claude Code is included with the Pro plan ($20/month). You can use it in the terminal, desktop app, VS Code, JetBrains, and on the web.
The main caveat is that Claude Code draws from the same usage limits as your regular Claude chat — so heavy coding sessions can eat into your quota faster. If you find yourself hitting limits often, the Max plans ($100–$200/month) offer significantly more headroom.
Thats a head scratcher:)
>$17 Per month with annual subscription discount ($200 billed up front). $20 if billed monthly.
Realistically the future of all this is that open models become good enough that LLM as a service becomes a commodity with a race to the bottom in terms of cost. Given where we are today I can easily see open weight models in 2-3 years making Anthropic and OpenAI irrelevant for everyday development work (I justify this like so: if my coding agent is 10x smarter than I am, how would I understand if it did all the right things? I want someone of roughly my intelligence for coding. I can see use cases for like independent pharma work or some such where supergenius level intelligence is justified, but for coding ability for mere mortals to reason about the code is probably more important).
You're really not going to miss CC. And OpenAI actually had some foresight to invest massively in compute so they don't run into usage and rate limits like Anthropic does constantly. I couldn't even use CC for more than a couple complex tasks before I was out of extra usage or session usage. It was a maddening productivity killer and I just switched to Codex full time.
After all, we may be a just a data source and not their intended demographic all along.
If Anthropic’s move is confirmed, my guess is other coding agents providers might end up making similar moves
Makes me curious about the internal thinking. One theory being they are in a capacity crisis and knocking Pro users off Claude Code is an emergency brake getting pulled. But an opposite theory is it's a revenue move and they think they have the lock in to pull it off. Especially if they are building up to IPO.
Interestingly the Team subscription which is still $20/month/seat still includes Claude Code. But you need minimum 5 seats. So it could be a way to force people off individual plans and into enterprise plans where possibly things scale better for them, especially IPO/wise. When one user wants it in a company, probably they go buy 5 seats.
My assumption is that people are able to very easily saturate Pro with Claude Code and therefore even though the quotas are lower (more than proportionally) the utilization of those quotas is higher enough that Pro is less profitable.
Of course, I don't have real insight into available compute, but the vibe slope seems to have dropped a bit, at the same time as new GPUs are being shoved into datacenters as fast as possible.
Loss of customers is the wrong direction
I dunno, I'm no business genius, but I think we're starting to see these companies try to find ways to make money instead of losing it.
Claude web is actually pretty good for dealing with random projects outside of code. I have a Home Assistant MCP server [1] behind a Cloudflare tunnel exposed to it that makes maintaining automations a lot easier.
“You asked, and we listened: Introducing Max Plus, our biggest plan yet, designed for those…” blah blah
I haven't changed the way I work, I've become more conscious of context sizes than before if anything.
Still run out of quota constantly with 4.7.
This guy's casual and crass response is a sign of disrespect for customers. Unfortunately, that is pervasive in the industry. The bubbles these teams work in are corrosive to empathy and real world impact.
The Claude constitution has a major section about not being deceptive. Now this is GTM, not the model, but there is clearly a coherence problem here... and if anyone should realize the important of their market positioning it's GTM.
Its funny that openai, who in my eyes went for the general public rather than devs initially, seems to be semi pivoting and catching all the fallout from anthropic's recent behavior.
It is a massive bummer, up until those few weeks ago, i was hard pulling for anthropic for quite some time, now i just dont care and hope something dope emerges quickly that signals i wont ever have to consider either of them.
While these tools stand to enable the democratization of productive capability in software engineering and other tasks (creating a renaissance for solopreneurs, let's say), what seems more likely to actually happen is that entrenched capital will become the only player with real access to this "knowledge as a utility" (was it Altman who called it that?).
We already see this playing out in two fronts: 1) the gradual reduction of services and 2) the DRAM market, where local-first tools (i.e., potential disruptors of the emerging "knowledge monopoly" created by the big AI firms) are being stifled by supply shortages. How many promising small-to-medium-sized competitors are being snuffed out of existence (or never starting) due to the insanity of the DRAM/storage/CPU (soon) markets?
The currently-subsidized access that we have to the big Opus-like models will, in parallel, be gradually be taken away until only the big players can afford it. And in the end what we will have is hyper-productive skeleton crews at a few consolidated firms performing (or selling expensive access to) basically all of the knowledge labor for society, with very little potential for disruption due to the hardware and "knowledge" scarcity engineered (in part, maybe) by this monopoly.
Not necessarily a closely held belief – just a hunch – which is why I want to see what parts of the picture I might be missing.
The real profitability is selling tokens to enterprise, and enterprise demand is growing so fast that they are short on the total amount of tokens they can generate per minute, and are prioritising rationally - enterprise gets a better experience - instead of optimizing for their lowest paying (and most loss leading) customers.
We are in a hardware crunch right now but that won't be forever, and eventually (likely 2028) we will get experiences like we got in January from pro-sumer accounts again.
It's easy to see this becoming a permanent position; the latest models and smarts are reserved for establishment members only, the riff-raff get the cast-offs. So the establishment is preserved and the status quo protected.
[0] I'm putting scare/irony quotes around this, but if the reporting is accurate, there is something to this; we built the internet on string and duct tape, it's not hard to see how a very smart AI could cut it to ribbons.
But there's competition out there -- the open-source chinese models. In their current form, I assume that will turn off many people but new models -- based on those -- are likely to appear. Also, OAI and Google will release new models and pick up the lost customers.
Now I'm going to learn more about local models. I don't need to be as good as a frontier model. Good enough and free from all this drama is a win for me
Most harnesses (claude, codex, opencode etc.) assume that you use a cloud model. There’s no sense of optimization or finer control.
Actually tried caveman mode yesterday and it made everything SO MUCH BETTER. GPT-5.4 has a habit of being extremely verbose to a ridiculous degree, it's like it's writing a report for a CTO or something and padding everything as much as possible to sound smart.
With caveman it just gives me lists of stuff in a compact format. Perfect.
Cursor was just acquired by SpaceX, so let's see what happens.
See [1] and [2] for an example of a support article that's had claude code removed as a Pro feature.
I guess this is the beginning of the end for subsidised model access, at least from Anthropic.
[1] https://support.claude.com/en/articles/8325606-what-is-the-p... [2] https://web.archive.org/web/20260420065828/https://support.c...
Wasn't this obvious from day 1 though? Can't see how anyone could've missed that.
They have now moved to be enterprise providers and don't need the cheap pro users as loss leaders anymore.
It makes no sense to do one of the higher tier plans unless they are directly generating you money.
Individual users barely matter. That's probably also the same group that decides to switch to Codex/Kimi/[whatever the hottest agent on any given day] on a whim, which Anthropic doesn't necessarily want to do business with.
Individuals are the ones that push for new tools at work though.
Source: what I witnessed at my company
If Anthropic is intent on losing the goodwill of the devs, they might not be happy with the consequences. Their product is quite commoditized at this point – the latest GPT, Gemini or GLM is just as good for most enterprise tasks.
- Rapid changes hurts the trust of your brand and product. In Google case, using a new service product became something you’ll think multiple times as you are more likely to axe it than rivals or specialized equivalents.
- While models currently has no clear winner. Anthropic’s core product is coding. But just as Skype, IE, Netscape their can always be another game changer you cannot count.
- The Pro plan is already limited for true agentinc workflows. The limits now are so bad that a business that relies on it would need bigger plans.
- Anthropic is already in a delicate situation where many devs are frustrated. Dropping or crippling the use even more just means this sector (which I can only assume is a big chunk) would switch to competitors tool that already try to compete.
- Local models, whether as Google sees it “edge” or even further would also take bigger part in the future.
Isn’t this the goal to some extent? They’ll probably have the standard “light” usage plan for weekend warriors or normal folk looking to play around. Companies that mandate usage and provide the subscription for hundreds of employees will have to cough it up, and will have no problem doing so if they want to compete with the others (or so the hype would allude to).
It is over for the little guy - home enthusiasts and vibe coders. Too many of them saturating resources for Max users.
IF you cannot afford few hundred dollars subscription go out and breathe fresh air. But if you can, watch where the ball is rolling - few thousand dollars subscriptions and even less programmers.
Something tells me congitively it's making us misjudge how productive it's making us.
It's clearly massively increasing output, but did the market already soak up all that productivity and now it's not compensated?
If your salary is 50k. And Claude makes you 2x as productive, why aren't you earning 100k?
Why is it anyone can't afford $200/mo if it's truely increasing worker productivity?
There seems to be a paradox here.
Personally I switched to Z.ai and GLM quite some time ago. I've not noticed any decrease in quality or quantity of my work.
Productivity wise employment is far more than code production productivity in a vacuum, and productivity gains are rarely captured by employees (see famous chart on worker productivity where that correlation changed around 1970). I wouldn’t expect to see much in the next 1-2 years besides noticing effective teams increasing velocity of features.
I think people in forums like complaining about things and aren’t representative of the broader set of people who are just using the tools, so no real paradox. For vast majority of tech jobs, $200/mo is still an absolute steal in terms of what these tools offer. Only the dullest of companies would not realize this.
Fwiw in the 80s-90s computers also didn’t really register in productivity metrics. Qualitative changes occur long before accurate measurement catches up.
Guess they’ve decided to move in the direction of allocating compute primarily to power users and enterprise.
But power users are not a sticky customer base. I just bought the ChatGPT Pro plan and would immediately switch over if the model performance is better and/or I get more compute.
Not too expensive
I heard they disabled signups for non-business accounts too.
Best forget about using Claude Opus models in Copilot.
Guess it democratizes it if you have money, huh?
(Head of Growth @AnthropicAI)
> When we launched Max a year ago, it didn't include Claude Code, Cowork didn't exist, and agents that run for hours weren't a thing. Max was designed for heavy chat usage, that's it.
Is there a wager that this is 100% foreshadowing Claude Code will be removed from the $100-200/month Max plans soon and go to something like API-only? Or only available on like a new $500-1,000/month plan? Restrict the $100-200/month ones to Claude.ai (website) or Claude desktop app only?
Either way, doesn't seem good to say it's a small test and then start justifying it in this direction.
Opus 4.6 is giving 2, maybe 3 questions before blowing through the Pro 5 hour limit as well. We are forced to use Sonnet which makes the same mistakes over and over and then to start trying with other companies. To make matters worse, it reuses old code as we try to survive between credit expiry so it re-introduced issues into the code with the limited credits, that we had already fixed on our own and with other models.
Anthropic in just a few days has gotten me to try GLM 5.1, the new Kimi, and back to OpenAI. OpenAI also seems to introduce new bugs without being carefully micromanaged. The advantage Claude has is that the models are more careful and can refactor code instead of leading to bloat as they go. But the throttling happening now is breaking things and making the entire subscription unusable. I really hope they fix it soon.
One interesting variable is that I'm located in Vietnam while my coworkers are located in Norway and Europe.
To work around this issue I used Claude for coding with a Copilot subscription which was much cheaper and had virtually no rate limiting.
Copilot gives you some set amount of credits each month, but you can also pay as you go if you run out of credit which is much better than the 5 hour window crap claude code would give me.
The only opus model available now on copilot for some reason is 4.7 and it costs 7.5x tokens, while everything else is 1x, 0.33x or free.
But I switched to using GPT 5.4 medium for a month or so which I find very reasonable.
At this rate I fully anticipate being able to run a comparable stack on a 128GB Mac Studio using quants of newer-generation distilled OSS models in a year or two. Being able to ramble to a computer for an hour about features and technical philosophy then have it build a nearly-working app for $50 is an exciting feeling. There's still a long tail of productionization and fixing what the model didn't adhere to but it's still incredible.
I got the 20$ gpt tier, and now i just use claude to craft MD plan docs instead, and then i hand them off to gpt 5.4 and it has been working great. can do about 4x as much work or so based on my feelings(not accurate). if i have just small simple stuff to do i might still fire those off with sonnet and that seems plenty viable, but as soon as its an opus tier task i swap to this workflow.
Little annoying as now im kinda trying to manage a .claude/ and an .opencode/ folder but i kinda just have the .opencode/ stuff reference the .claude/ stuff so its a little less bleh.
I've been keeping within my usage because ive been in a funk a bit, but when i was slightly more worried id sorta just juggle whether claude or gpt would handle writing some initial tests as it did seem to kinda be imbalanced otherwise. seems like gpt just spam resets weekly usage throughout the week anyway so its prolly nbd.
Glad I’m not the only one!
I’ve been limited so often this week I’ve setup half a dozen token compression tools in my workflow and had to do a crash course in token optimization.
Of course, it seems to only slightly delay the inevitable and doesn’t really solve the problem.
There is a lot of political capital to be earned by appearing to be "tough" on AI companies.
One thing is clear, Anthropics communications and leadership is horrible. You don't launch or remove features like this. How this is communicated and handle is something like mom+pop shop would do.
Existing subscriptions are not impacted according to Tweets from their team. It’s apparently an A/B test they’re rolling out.
If you actually wanted the $20/month Claude Code plan you may have just shot yourself in the foot.
It is pure speculation of course, but I don't have any other explanations on the stuff they are pulling in the last 2 months.
Would not be surprised to see OpenAI follow suit.
Or perhaps OpenAI's LLMs are just so more compute efficient that they can actually offer that sustainably...
"Codex will continue to be available both in the FREE and PLUS ($20) plans. We have the compute and efficient models to support it."
Both the compute, and the efficient models.
>We will do this every million users up to 10 million.
>Happy building!
https://x.com/sama/status/2041658719839383945
Last reset today, after the 4 million users milestone.
These shenanigans are earning them no respect. The market is already annoyed on model serving QA issues, and now (recently) Opus limits. They don't want to lose to OpenAI - understandable - but these shortcuts won't earn them anything either.
I have an unlimited-usage API billing plan through my dayjob, but for obvious reasons they don't allow piggybacking personal usage onto that. so I paid for the $20/mo personal plan as an easy and relatively cheap method of professional development / keeping my skills current. I don't particularly mind paying $20/mo, but I'm absolutely not paying $100/mo.
also, part of the reason I didn't mind paying for the personal subscription is that I liked having consistency between the tools I use for my dayjob and the ones I use for side projects. if that goes away, then I might as well switch away from Claude usage at work as well. I very much doubt Anthropic's revenue predictions for this change are taking things like that into account.
making a change like this without an announcement, just sneaky updates to product pages, is also an absolutely unforgivable thing to do, in terms of me trusting them as a company.
I think the only reason to do this would be that they just can't scale up to service the volume they have and need to cut down significantly on the total number of users. Seems also like a rough business proposition. Most of the pro plan users would probably migrate to a competitor at a similar price point (I know I will).
The only other possibility would be if they are losing too much money on the compute power and just can't offer it at that price anymore. But then upgrading the plan gives you more compute per dollar, so maybe they're just banking on people not actually using all of what they pay for?
I had previously thought that the inference cost of using a trained model was relatively low and that most costs went into training new models, but maybe that is less true with the more powerful newer models.
If it costs a ton more to serve Opus vs serving something like Kimi or Qwen, then I think most people just won't use the more expensive version for most things.
However, my company paid for my annual subscription, so maybe I'll ask our lawyers for advice - the only reason they paid for this was my access to CC and with my use the next tier wouldn't make sense, AND no one will expect Anthropic to not nerf it too.
I would not jump to conclusions yet.
https://web.archive.org/web/20260421141017/https://claude.co...
Edit: fixed the url thanks to scq
Might have been taken down?
Let’s say my trust level and appreciation for the product for the past month had a big negative hit for me
It would seem misleading to sell monthly, or even yearly, subscriptions under the guise Claude Code comes with the subscription, for it to only be yanked out underneath you. (Although depending who you ask, Anthropic have already done actions similar to this).
If they rugpull Claude code from my already paid for annual subscription I’ll have to issue a chargeback
Anthropic better get this sorted out with a proper product manager and marketing or they risk customers jumping to easier to understand platforms that are good enough.
https://www.anthropic.com/news/anthropic-raises-30-billion-s...
I reckon they'll be fine. Not agreeing or disagreeing with you, but they have enough customers who won't leave.
Otherwise companies will keep exploiting using their dominant position.
We were taught dictators are bad, monopolies are bad, but now allowing 2 companies control most of the software development
but why not work towards it?
* elect politicians who will support this
* change laws to accommodate it - if corporations train data, on every usage they should pay higher taxes so they can't exploit the open data, but public ledger trained model is fine to use open data
* similar tech exists (bitcoin, torrent), needs some modifications
On the one hand, the people there are supposedly among the smartest on the planet. On the other hand, they consistently forget that they're dealing with LOYAL humans, and these humans prefer respectful communication beforehand instead of being messed with every other day.
My hope for reasonable behavior is to not handle it this way. Decrease limits and increase prices if you can't handle it and be _honest_ about it.
Are they just looking for a way to rationalize another hostile act? And already have expectations like:
- "minus 10% in pro signups" -> oh, let's drop those coders who won't pay anyway
- "minus X% in pro signups and plus X% in max" -> awesome, PAY UP!
In general Anthropic seems to be pretty bad at clearly communicating what is going on. I have both Claude Pro for Claude Code and ChatGPT Plus for Codex, and lately I've been reaching for Codex first more and more often... at least for the hobby stuff I'm using Claude/Codex on, they seem pretty much equivalent in terms of practical capability/usefulness.
Should we instead use a generic coding agent with a particular model and just pay per token?
So I pay for Codex instead.
Why not with email?
Would it really be that hard for them to just make all of the changes and then do a redeploy rather than doing them incrementally? It's not like they're just editing the raw HTML sitting on the server manually, right? Actually, don't answer that, I'm not sure I even want to know the answer.
3 hours later…
Then some genius intern will say "if we offer it to the lowest level plans, the users will get hooked on how awesome it is!"
Then Anthropic will put it back.
Or they could just ask Claude if it is a good idea to remove.
Additionally I run a constant hacking contest between GPT and Claude. It’s a toy project and it simulates an attack/defense of a small corporate network.
Claude used to win pretty handily. Suddenly it’s started to lose 90% of the time. I thought GPT had gotten better but no, looking at the logs it seems that Claude is slower and more prone to running in circles. This is still the case when switching to Opus 4.7.
I don’t know what that means but it’s undoubtedly worse.
If Anthropic continues to getting worse, try Amazon Kiro and other companies that run Claude on their own hardware.
It might be expensive and have a worse experience compared to Claude's code, but at least the model itself is the "original flavor."
These days, it's hard to ask for much.
I could be connecting unrelated dots here, but it sure as hell seems quite coincidental to me.
I realize this duplicates a lot of sentiment already in this thread but anyone here with pull at Anthropic please understand it will undo a lot of the goodwill that made Claude so successful in the first place.
I settled for the AMD rough equivalent. It’s not perfect but it can still handle most of the work. Now if only extra ram would come down in price… I find I need about 5 GB more than I have
From what I can tell Opus 4.7 is more resource-intensive than Opus 4.6 is more resource-intensive than Opus 4.5.
> For clarity, we're running a small test on ~2% of new prosumer signups. Existing Pro and Max subscribers aren't affected.
April: "The fact that we're doing X isn't news because we're only starting to do X"
August: "The fact that we've fully rolled out X isn't news because we started X in April"
> Engagement per subscriber is way up. We've made small adjustments along the way (weekly caps, tighter limits at peak), but usage has changed a lot and our current plans weren't built for this.
https://xcancel.com/TheAmolAvasare/status/204672528250217304...
Even the downtime would've been fine (as GitHub shows). Instead they're pissing it all away by letting employees make random announcements on random platforms.
I remember when they first added Claude Code to Pro — it was limited to Max initially — and my first thought was that it seemed kind of stupid, because at one fifth of my current limit, I would be hitting walls all the time...
But I’ve mostly been using it for gitops infrastructure in my homelab. I wonder if the token usage is lighter than if I were developing an application.
OpenCode and their Go plan will get you close if you're willing to put in the config work.
For when you do need the larger models Fireworks has a pretty generous 'Pass' that comes out to about $7 a week for some of the larger bleeding edge models.
Other than that Codex's $20 plan is still somewhat valuable though they keep reducing usage. Google's $20 plan will get you some Opus usage in Antigravity and a generous amount of Gemini. Not sure how long that will last as they've been tweaking pricing and planning language recently too.
Opus 4.7 consumes tokens at a faster rate and folks were complaining that the Pro plan included too few credits for real work.
And Anthropic now allows `claude -p` (which invokes Claude code) for 3rd party agents like OpenClaw, which consume far more tokens by running autonomously, 24/7.
https://bsky.app/profile/mattgreenrocks.bsky.social/post/3mk...
Another example, I recently saw two people over on Twitter posting LLM responses at each other in a bitter argument about Vercel's security breach. They made no attempt to pretend they'd formulated the ripostes themselves, it was just screenshotting one-sided conversations... What's the point? They could've saved themselves the trouble by spawning two LLMs, naming them "John Doe" and "Fred Doe", then telling them to argue and post the name of the winner.
Disclaimer: I don't use Twitter, Bluesky, Mastodon, etc., so maybe it's not that deep.
Opus is fairly useless on Pro given the rate limits anyways.
Folks are assuming that only the $100 plan will include Claude Code access. I think a more likely scenario is that everyone will be able to use CC.
Now though I don't dare use spend tokens for basic note taking with Sonnet because I'm hitting the limit over a couple million tokens on the 20x plan, so they've really tightened the purse strings since November.
The Anthropic website has become inconsistent. Some places say Claude Code is included in the Pro plan, other pages don't.
Others in non-tech sectors are forced to use Copilot. Who knows what I would pay for a usable LLM out of my own pocket. Probably more than $200.
That is the only way to avoid being held captive by Anthropic / Meta / Google.
Business accounts are like max 6x accounts.
But it seems this is all in a state of flux.
And there’s the lovely asterisk at the bottom:
> Prices and plans are subject to change at Anthropic's discretion.
The million token context + reduced caching period + new models using more tokens made this a probably unpopular but perhaps unavoidable development.
There's a hard problem here balancing costs and experience. I'm afraid despite the bad experience for people that this is necessary and $20/month was just too big a loss to sustain.
Is there any marginal cost associated with a new subscriber?
I have always heard inference is cheap and the cost was in training, so I assumed any subscriber was making them money, just not enough to cover their insane fixed costs.
But I am just guessing.
I strongly suspect both are wrong.
I got rate limited after about 30mins of coding and was thinking, who the hell i going to work like this?
So they really seem to be running into extreme capacity issued now.
The folks hurt most by this are serious people in developing countries and young people starting out. Occasionally a dabbler turns into a serious user but I imagine that’s far less likely than people wish it were.
The value to companies who make these changes is they don’t have low value users or low value contributions to worry about, which has its own not insignificant overhead. In the age of AI slop everywhere we’re likely to see a lot more attempts to separate the wheat from the chaff.
Assuming this limitation applies to team seats in the same way, at $20/mo, businesses could afford to have everybody on the plan. Plenty of folks write only a few hours of code per day—or even per week in their job. These are still professionals, not dabblers.
It’s the Apple model. Yes you pay a ton more. But my 2013 MacBook Pro 15 I got in college lasted 10 years and was still fine even when it was stolen. That’s what you pay for. You pay for a ton of built in apps and functionality and quality.
Arbitrarily removing things is customer and more importantantly good will hostile.
That $20/month is not profitable? That Anthropic thinks that people are willing to pay a 400% markup without batting an eye? That Anthropic is desperately trying to clean up their burn rate? Why should we trust a company that can screw up basic PR this hard?
As someone who tries to manage usage for a small team they just added Claude Code to the Standard Team seat now they are removing it!?
Not to mention that they will ban your entire organization from a bot deciding you violated their TOS with no communication and no way to contact anyone to understand what happened.
If this is real we are switching to OpenAI or Gemini it is not worth all this non sense
That sucks, I guess I'll cancel my Claude account. Not paying 100 dollars. That's crazy
Me: Why is your website saying Claude Code is not included in Claude Pro subscriptions?
"Fin", the slow, dopey AI agent: That’s confusing, but it usually comes down to what “Claude Code” means in that context. Claude Pro does include Claude Code (the Code tab in the desktop app, and Claude Code access for individuals).
What Pro does not include is API access via the Claude Console. Those are separate products: Pro is for chat, and the Console is for building apps with the API.
So if the website is saying “not included,” it’s likely referring to Console/API access, not the Claude Code feature itself.
Did that clear up the confusion?
Me: No. It clearly states on the pricing page that Claude Code is not included.
Fin: That’s frustrating, <NAME>. The key detail is that Claude Code is included in Claude Pro, but API access via the Claude Console is not.
So if the page you’re looking at says “Claude Code is not included,” it’s almost certainly referring to Console/API access (the developer platform), not the Claude Code feature itself.
If you tell me the exact URL or screenshot of the page that says that, I can confirm which “Code” it’s talking about.
Was that explanation helpful?
Me: https://claude.com/pricing
Fin: On https://claude.com/pricing, Pro does include Claude Code. It’s listed right there under Pro’s “Everything in Free, plus” bullets: “Includes Claude Code.” So if you’re seeing “not included” somewhere else on that page, it’s likely referring to something different than the Claude Code feature. The one thing Pro does not include is API access via the Claude Console (that’s a separate product).
Also, the Code tab in the desktop app requires a Pro (or Max/Team/Enterprise) subscription, and it may not show for everyone yet since it’s still rolling out.
Does that clear things up for you?
Me: there is clearly an "x" at this page. escalate to a human immediately.
and then the cycle continues. I'm cancelling my Anthropic subscription, sick of worrying about limits and getting banned and simply awful experience.
However, if you think you'll need the extra capacity of Max, it's $100, not $200.
I thought we now had advanced tools to which we could ask to do things like: "Remove all mention of Claude Code in the Pro, but not in the Pro Max plan".
But apparently the CGI-days called and asked the webmaster to manually edit .html files one by one?
Maybe this is coming next
"We've determined that claude code is too dangerous to your code base to release, so we are withdrawing it"
I know, crazy idea. When we told you they’re getting you hooked and would rug pull you called us permanent underclass or something.
Here's my hot take: Anthropic et al. are trying to make developing a subscription-only job, and they've done that by illegally pirating pretty much the whole Internet. If they were to go out of business tomorrow and serving models was to become a commoditized service like storage we'd be all better off. Sure, we would have less research on frontier models, but we don't need AGI, we need good local models, RAM and good open source / weight AI tools.
Since then, I had to add:
"or won't let you log in?": https://github.com/anthropics/claude-code/issues/44257
"or makes stuff up?": https://dwyer.co.za/static/claude-mixes-up-who-said-what-and...
"or when it's down?": https://status.claude.com/incidents/6jd2m42f8mld
"or when you get banned?": https://bannedbyanthropic.com/
"or installs spyware?": https://www.thatprivacyguy.com/blog/anthropic-spyware/
And this is all exclusively about Anthropic. It's insane. On any other tech, there would be a consensus to wait until it's stable, but not AI - we go full throttle when it's AI.
Genuinely curious how people who have implemented this in serious companies are answering these questions, because my answer is to keep it the fuck out.
This makes me think either they’re severely resource constrained and need to focus on “high value” customers, they’re bleeding money on inference, or their sales and marketing team is incompetent.
Regardless, this feels like a pretty big rug pull. Especially without a phase-out period and a real announcement. As someone using Claude Code on a personal hobby project to get a better feel for its capabilities, I’m not sure what to do now. I can’t justify the $100+/mo plans for a hobby project.
My choices are then:
- Code this project by hand, which would be fun but defeats the point of this being my agentic coding project.
- Find another model and use Codex or OpenCode or whatever.
- Put the project on a shelf till this shakes out.
This was never the case though. There's a per week and per 5 hour quota. If you exhaust either you have to wait for the reset. What they're doing makes no sense.
https://claude.ai/share/1a4293bd-b2d4-41b7-a887-eb42b3ae8b6e
“ The standard answer here is no — Anthropic does not typically refund the unused portion of annual plans , and annual subscribers won’t see prorated refunds, retaining access for the full remaining period instead. That said, your situation is a bit different — you’re not just canceling, you’re canceling because a feature you paid for was removed. That’s worth contacting Anthropic support directly about. Their support team can check your refund eligibility , and this kind of material change to the plan is exactly the case where a support escalation could go differently than a standard cancellation. You can reach them through the in-app support messenger at support.claude.com or via the thumbs-down feedback button. I’d recommend explaining specifically that Claude Code was a factor in your annual plan purchase. ”