[HN Gopher] Claude's Max Plan
       ___________________________________________________________________
        
       Claude's Max Plan
        
       Author : nanfinitum
       Score  : 112 points
       Date   : 2025-04-09 18:07 UTC (4 hours ago)
        
 (HTM) web link (www.anthropic.com)
 (TXT) w3m dump (www.anthropic.com)
        
       | serjester wrote:
       | Bold to launch this before the roll out their own reasoning
       | models / deep research. Seems like table stakes if you want to
       | capture power users but maybe that's just my workflow.
        
         | oidar wrote:
         | They already have a reasoning model out. I don't think deep
         | research is on the agenda atm.
        
         | HarHarVeryFunny wrote:
         | Anthropic have recently said that they don't like the
         | "reasoning" label and don't intend to have separate base and
         | reasoning models, but rather one model that does it all. The
         | current Claude 3.7 already seems to be doing some reasoning,
         | displaying "pondering" and "analyzing" messages between
         | different stages of output generation.
        
           | logicchains wrote:
           | Regardless of naming, they don't have a model competitive
           | with o1 Pro for coding (or with Gemini 2.5, for that matter).
           | As someone who pays $200/month for GPT Pro I definitely
           | wouldn't pay the same for Claude (and am considering
           | cancelling GPT Pro if they don't deliver something better
           | than Gemini 2.5 soon).
        
             | CamperBob2 wrote:
             | Gemini 2.5 is awesome, but when it screws the pooch, it
             | _really_ screws the pooch (e.g.,
             | https://gemini.google.com/share/374ac006497d ). It has
             | delivered some of the best responses I've seen lately but
             | also some of the worst.
             | 
             | I have been tempted several times to kill my GPT Pro
             | account, but it is still valuable for cases where Gemini
             | and Claude don't get the job done for whatever reason.
        
             | tesch1 wrote:
             | I tried gemini 2.5 just once for coding and it immediately
             | spit out hilariously invalid python, not even clearing the
             | lowest of bars.
             | 
             | What (language/topic) are you coding in that gemini (or
             | even chatgpt) is better than claude? Very surprised to hear
             | this.
        
               | logicchains wrote:
               | For Golang and Haskell I found o1 Pro and Gemini 2.5 are
               | much better at generating code that works first try than
               | Claude. Gemini 2.5 in particular can generate thousands
               | and thousands of lines of code that correctly use types
               | and functions that were generated earlier in the context,
               | with minimal errors.
        
               | shmoogy wrote:
               | I've been using Gemini pro 2.5 over Claude 3.5 in cursor
               | all week - some people have Gemini do a prd and tasks and
               | Claude do the actual coding. I switch between them but
               | have been pretty impressed with Gemini.
               | 
               | Gemini sometimes fucks up diffs or doesn't actually apply
               | the edits - Claude is rock solid at that. They're both
               | very good though - but 3.7 really likes refactoring
               | unneeded shit and removing code sometimes.
        
         | jasonjmcghee wrote:
         | Their reasoning is called "extending thinking". It was released
         | alongside Claude Sonnet 3.7.
         | 
         | They have web search, but it's true, no "deep research". It
         | honestly is not very good and it's WAY too trigger-happy with
         | it. As a result, if you accidentally leave it on, you get
         | terrible answers to simple questions that non-search mode would
         | have answered well.
         | 
         | (And for context Claude Sonnet 3.7 is my model of choice)
        
       | smallerfish wrote:
       | At last. I honestly think the product engineering org at
       | Anthropic has serious issues. They are slow to ship, and quality
       | of the webapp has been decidedly mixed in the past 6 months. Some
       | of the UX decisions they've made are bone-headed.
       | 
       | (OpenAI launched their pro plan 4 months ago!)
        
         | bastardoperator wrote:
         | I just assumed most people at this point are using a client.
        
           | tener wrote:
           | It is hardly better though. Aside from MCP support (with
           | rather poor UX, just compare it with anything else out there)
           | it really has nothing to offer.
        
             | bastardoperator wrote:
             | I've been satisfied with msty and code/cursor.
        
         | yieldcrv wrote:
         | I actually like Claude over ChatGPT specifically because of how
         | their web UI distinguishes.
         | 
         | Its clear that neither company prioritizes their flagship chat
         | UI, over their other offerings. So if it will be in stasis, I
         | definitely like it
         | 
         | There are other web AI offerings now (deepseek web, perplexity,
         | to name a few). This is what I like between these two:
         | 
         | The speed of responses.
         | 
         | The reliability. ChatGPT web UI often times out or has server
         | errors for random things.
         | 
         | Different censorship. Overall better for my use cases, and how
         | errors are handled.
         | 
         | Initially search of old chats was a first class citizen in
         | Claude. ChatGPT didnt have it, needed third party plugins, it
         | has since updated, but is still lackluster in this regard.
         | 
         | The diagrams, charting and coding interface is great and
         | better. Easier for seeing how something compiles and runs,
         | easier to copy and paste.
        
       | jtwaleson wrote:
       | A bit off-topic, but I wonder when Cursor is going to see a
       | massive price increase. I've tried Aider (granted, this was when
       | GPT4 prices were still much higher) and spent $10 in one hour
       | easily. Now I use cursor a LOT (100h focussed programming time
       | per month) and mostly stay within my $20 monthly fee. I think
       | they must be losing money on customers like me.
        
         | ZeroCool2u wrote:
         | Have you tried Roo Code or Cline instead? I never felt the need
         | to use Cursor after trying them and they're just extensions, so
         | it's easy to install and use your own API key.
        
           | jtwaleson wrote:
           | How bad is the bill?
        
             | seunosewa wrote:
             | With Gemini 2.5 the bill is zero for now.
        
         | danialtz wrote:
         | it is already super expensive if you want to get an actual work
         | done using their max model. On a midsize project, it cost ca.
         | EUR200 for my vibe coding tests to get something reasonable
         | done (each call, each tool costs 0.05c). Their normal claude
         | window is super short, and almost unusable for serious work.
         | Stack was python and Nuxt.
        
           | SkyPuncher wrote:
           | > to get an actual work done using their max model
           | 
           | I get plenty of work done without the max model.
           | 
           | The key is breaking out a planning step before
           | implementation.
        
       | podgorniy wrote:
       | Lol. The best they can propose for 200 is "not throttling"
        
       | CosmicShadow wrote:
       | I already had to go through so many hoops to sign up for a teams
       | account for 5 separate accounts for myself, I wonder if this will
       | be any better or not. At least if one account starts getting dumb
       | or somehow runs out of context, I can just switch, but if
       | something goes wrong and your $200 account exponentially loses
       | context, then you are stuck waiting. Anyone have any idea of the
       | actual differences?
        
       | sharkjacobs wrote:
       | I use the free model and API, when I look at the pricing page I
       | see (for the extant $20 Pro plan)                   - More usage
       | - Access to Projects to organize chats and documents         -
       | Ability to use more Claude models         - Extended thinking for
       | complex work
       | 
       | What does "more usage" mean? It doesn't say anywhere what the
       | free tier usage limits are. What are "more" models? It also
       | doesn't make clear what models are available with each tier
       | (except for "extended thinking" which is a separate bullet point)
       | 
       | The only thing I can reason is that they want to keep this vague
       | so that they don't have to update their marketing copy each time
       | they update their offerings, but that's absurd.
        
         | mathgeek wrote:
         | Not sure if it still holds, but it was 5x at launch:
         | https://www.anthropic.com/news/claude-pro
         | 
         | FAQ agrees:
         | https://support.anthropic.com/en/articles/8324991-about-clau...
        
           | mvdtnz wrote:
           | 5x what though? Even if we assume this is the true figure,
           | everyone who uses Claude regularly knows that usage limits
           | fluctuate over the course of days and weeks.
        
             | section_me wrote:
             | Surge pricing? I'm not sure if my remark is sarcasm or a
             | prediction on the future.
        
               | TheOtherHobbes wrote:
               | In the near future, everyone will have an intelligence
               | meter in their hall and pay IQCoin to the National
               | Intelligence Utility based on how much intelligence they
               | use each month.
        
             | fragmede wrote:
             | not just limits, but response times. I swear there's a Cron
             | job or something that kicks in a 2300 pacific because
             | Claude backs up right around then.
        
         | jsheard wrote:
         | > It doesn't say anywhere what the free tier usage limits are.
         | 
         | Ah, I see they've been to the Cloudflare school of free tier
         | bait-and-switching.
        
           | pdyc wrote:
           | can you elaborate you mean CF workers or the pro plans?
        
             | jsheard wrote:
             | I mean how CF has advertises "unlimited bandwidth" and
             | "unmetered DDOS protection" on their free and flat-rate
             | paid plans, which abruptly turns into not-unlimited pay-
             | per-gigabyte if they decide you're using them too much, but
             | you're not allowed to know where that line is so you can
             | plan around it. It's a fun surprise where they suddenly ask
             | you to pay 10x more out of nowhere.
        
               | tomnipotent wrote:
               | Can you point to a single example of CF doing this to a
               | legitimate entity that wasn't obviously abusing their
               | plan? Every time this comes up it's always some shady
               | website that's getting the limited treatment.
               | 
               | And it's not "unlimited bandwidth", it's "unlimited
               | bandwidth with restrictions". That's an important and
               | critical distinction. I'm not sure I understand the
               | position that somehow CF needs to subsidize every
               | customer.
        
               | webstrand wrote:
               | Nobody said CF needs to subsidize every customer. The
               | rules need to be clear though, not secret. Same with
               | anthropic's usage limits, since they're not public they
               | can degrade the service you're paying for at any point in
               | time with no notice.
        
               | Veen wrote:
               | The oxymoronic style of marketing bullshit: unlimited
               | with restrictions, free but only if you pay, lifetime
               | membership for three years, etc.
        
               | brutal_chaos_ wrote:
               | Just a tiny nit, 10x more of 0 is still 0. Your point
               | still makes sense though
        
         | miki123211 wrote:
         | "more" usage likely means that they have a limited number of
         | GPUs, and what models you get access to depends on how much
         | you've used them recently, but also on how busy the GPUs are at
         | this moment.
         | 
         | This is also how batching works for API users. If you don't
         | need the results immediately, you can give them a batch with an
         | attached 24-hour deadline, and they'll slot you in whenever
         | they expect low usage, in exchange for better prices.
        
       | mvdtnz wrote:
       | I have been a heavy user of Claude but cancelled my Pro
       | subscription yesterday. The usage limits have been quietly
       | tightened up like crazy recently and 3.7 is certainly feeling
       | dumber lately.
       | 
       | But the main reason I quit is the constant downtime. Their status
       | page[0] is like a Christmas tree but even that only tells half
       | the story - the number of times I have input a query only to have
       | Claude sit, think for a while then stop and return nothing as if
       | I had never submitted at all is getting ridiculous. I refuse to
       | pay for this kind of reliability.
       | 
       | [0] https://status.anthropic.com/
        
         | bakugo wrote:
         | Can confirm. As an API user, "overloaded" errors have been
         | happening pretty often lately.
        
         | tesch1 wrote:
         | At 99.38+ uptime - seems like the reds in the xmas tree are
         | OR'd into the bars rather averaged in, making them look worse
         | than they actually are, which is refreshingly honest to see in
         | an uptime monitor.
        
         | bayarearefugee wrote:
         | Also cancelled Claude Pro due to a combination of unreasonable
         | amounts of downtime and Gemini 2.5 just doing a better job for
         | me.
         | 
         | I'd estimate there's probably like another 6 months to a year
         | where I'll jump around to whatever 'cloud' LLM is currently
         | winning on quality of the model (in terms of usefulness as a
         | coding assistant) and just basic UI/availability and then I'll
         | build a locally hosted system and just use that.
         | 
         | I certainly don't have anything like brand loyalty to any of
         | them, so I'm down for a race to the bottom.
        
       | dzhiurgis wrote:
       | They are a bit of ripoff, especially cli thing. It will
       | hallucinate tons of things, add things that break code and then
       | you pay again (quite a lot) to fix it.
        
         | pphysch wrote:
         | Sounds like a great business model
        
           | dzhiurgis wrote:
           | Yep. Being inconsistent makes it even more addictive than
           | being perfect every time.
        
       | brunoqc wrote:
       | The email subject line was "We heard you: Introducing the new Max
       | plan for more Claude usage".
       | 
       | What the hell?
        
         | Mystery-Machine wrote:
         | Can you elaborate which parts are frustrating to you? It wasn't
         | clear who the email came from just from the subject?
        
           | brunoqc wrote:
           | Sorry. The "we heard you" and then offered a ridiculous price
           | tier.
        
             | ZeroTalent wrote:
             | People were complaining about rate limits on the $20 plan.
             | 
             | How is $200/month ridiculous for groundbreaking technology?
        
       | prompt_overflow wrote:
       | F
       | 
       | I can't believe we're going to high pricing based subscription.
       | This makes me think we need than ever open source models like
       | qwen and deepseek.
        
         | yahoozoo wrote:
         | It's because these companies are burning billions of dollars.
        
         | Pikamander2 wrote:
         | On the other hand, the free options are better than ever. I've
         | only used the free versions of ChatGPT, Claude, and ImageFX and
         | have been impressed by how much you get without spending a
         | dime. The only real limitation I regularly feel is not being
         | able to upload something like a 1GB CSV file for analysis, but
         | I guess that's a fair restriction for a free web tool.
        
           | johnisgood wrote:
           | I am impressed by the free version of ChatGPT, not so much
           | with Claude's (claude.ai). I love Claude Pro, more so than
           | the paid version of GPT, but damn, the free version of Claude
           | is awful, you reach the limit extremely quickly. The free
           | version of ChatGPT works quite well.
        
         | int_19h wrote:
         | The problem is that comparable-quality open models will still
         | require lots of hardware to run at reasonable speeds, so even
         | then it's not cheap.
        
       | woadwarrior01 wrote:
       | Their pricing page says $100/month
        
         | beklein wrote:
         | - Expanded Usage: 5x more usage than Pro => $100 per month
         | 
         | - Maximum Flexibility: 20x more usage than Pro => $200 per
         | month
        
       | ozten wrote:
       | I want different capabilities at $200.
       | 
       | I paid for that to get access to Deep Research from OpenAI and I
       | feel I got more than $200 of value back out.
       | 
       | These companies have a hard time communicating value.
       | Capabilities make that easier for me to understand. Rate-limiting
       | and outages don't.
        
         | matwood wrote:
         | I waited until Deep Research came to the normal paid plans, but
         | it's been very useful the times I have thought to use it.
        
         | tibbon wrote:
         | My current dream is a model that's good at coding with a ~10m
         | token content window. I understand Llama 4 has a window
         | approximately that size, but I'm hearing mixed results on its
         | coding capacity.
         | 
         | If it had deep research and this, with a large number of API
         | requests, I'd consider $200/month.
        
           | ZeroCool2u wrote:
           | I mean you get a 2 Million token context window and by far my
           | favorite coding model with Gemini 2.5 Pro.
        
             | siva7 wrote:
             | Has someone tried the 2m context window for a code base and
             | can report how it compares over claude or o1?
        
               | ZeroCool2u wrote:
               | I mean I've tried it with Gemini 2.5 Pro + Roo and then
               | tried Claude 3.7 + Roo on the same task and Gemini blew
               | Claude away. Haven't spent anymore OpenRouter credits,
               | because Gemini was so much better.
        
               | johnisgood wrote:
               | Does Gemini have a web interface similar to claude.ai? I
               | am lazy[1], but I am also poor. I would not be able to
               | afford 100 USD per month.
               | 
               | [1] But if it is cheap enough, has large context window,
               | then I might consider setting up something akin to
               | claude.ai with Gemini's API.
        
               | ZeroCool2u wrote:
               | Yeah AI Studio is free with decent rate limits, though
               | obviously more developer focused:
               | https://aistudio.google.com/
               | 
               | The official Gemini app works well for me too and there's
               | a nice free tier and it's free if you have a newer Pixel
               | phone. Otherwise $20/month for the Advanced tier. There's
               | no $200/month option.
               | 
               | https://gemini.google.com/app
        
               | johnisgood wrote:
               | I mostly use LLMs on PC, as I use LLMs mainly for coding.
               | 
               | Does AI Studio allow you to have projects with project
               | files and whatnot?
               | 
               | How about its context window length, more or less than
               | Claude's?
               | 
               | I am also interested in open-source alternatives to the
               | web interface that claude.ai has, I know there are some
               | but I have forgotten their names, would be cool to have a
               | list here.
        
               | simonw wrote:
               | The best open source UI I know of is
               | https://openwebui.com/ - you can point it at any OpenAI
               | API compatible endpoint and both Gemini and Anthropic
               | offer those now.
               | 
               | You can use the Gemini API for free with quite generous
               | allowances, including for 2.5 Pro.
        
               | johnisgood wrote:
               | Thanks Simon, will take a look.
               | 
               | Extremely off-topic: are you still around DS?
        
               | simonw wrote:
               | DS?
        
               | indigodaddy wrote:
               | There's also Google's https://idx.dev - which is a
               | webide/vscode dealio and you can use gemini in agentic
               | mode (mix of 2.0/2.5 but if you use your own gemini key
               | you can guarantee 2.5 Pro i think)
               | 
               | edit, well it now appears to be https://firebase.studio/
               | - that is a recent change I haven't used it since it
               | changed its name..
        
               | bionhoward wrote:
               | AI studio is only developer focused if you're not working
               | on AI, which is a prohibited use case according to the
               | Gemini API / AI Studio "Additional Terms"
        
               | khromov wrote:
               | Made a video comparing Gemini 2.5 Pro to Claude Sonnet
               | 3.7 recently: https://www.youtube.com/watch?v=AVdVJ_hD_vo
        
             | sva_ wrote:
             | I just subscribed to the free trial yesterday, and I've
             | been hooked tbh. I haven't subscribed to any of the other
             | LLM companies so far. I hope something else comes out
             | within a month because I really don't want to spend 22 Euro
             | per month for it.
             | 
             | The 1M context window (2M?) really sets it apart.
        
               | baq wrote:
               | 22 euro per month is less than 1 per day. Less than one
               | espresso.
               | 
               | I get the subscription fatigue, but there are splurges
               | and there are truly valuable things.
        
               | simonw wrote:
               | I believe you can still use Gemini 2.5 Pro for free via
               | https://aistudio.google.com and their gemini-2.5-pro-
               | exp-03-25 model ID through their API.
               | 
               | The free tier is "used to improve our products", the paid
               | tier is not.
        
           | vessenes wrote:
           | The results are not mixed, Llama 4 is terrible at coding. I
           | agree on longer context window being the dream.
        
           | imiric wrote:
           | Has anyone found the output at these large context windows
           | usable at all?
           | 
           | IME the quality of all models goes down considerably after
           | just a few thousand tokens. The chances of hallucinating,
           | mixing up prompts, forgetting previous prompts, etc., are
           | much more likely as context size increases. I couldn't
           | imagine a context of 1M tokens, let alone 10M, being usable
           | at all. Not to mention that any query is going to come to a
           | crawl just to move that amount of data around (which still
           | annoyingly happens on every query...).
           | 
           | So usually at around 10K tokens I ask it to summarize what
           | was discussed, or I manually trim down the current state, and
           | start a new fresh chat from there. I've found this to work
           | much better than wasting my time fighting bad output. This is
           | also cheaper if you're on a metered plan (OpenRouter, etc.).
        
       | mohsen1 wrote:
       | I cancelled my OpenAI plan. Gemini 2.5 Pro is extremely good
       | compared to OpenAI and Anthropic models. Unless things change, I
       | don't see why I need to pay those subscription fees?
        
         | nanfinitum wrote:
         | Yeah, I'm not really sure what the long play is here. $200 is
         | what I spend on groceries for an entire month.
        
           | paulddraper wrote:
           | I'm impressed.
           | 
           | A pound of chicken breast, a pound of apples, a third loaf of
           | bread cost at least $7. And that's only 1500 kcal.
        
             | yzydserd wrote:
             | Probably spends another $600 on takeout.
        
               | ZeroTalent wrote:
               | and $1,000 on APIs :D
        
             | fragmede wrote:
             | COL isn't the same everywhere. That $8 of chicken in
             | downtown San Francisco whole foods that closed is $4
             | elsewhere and those differences add up.
        
             | throwup238 wrote:
             | A ten pound bag of russet potatoes costs $2-3 here (high
             | CoL SoCal), and that's >3,000 kcal. A four pound bag of
             | pinto beans is $4, that's >5,000 kcal. That's four days of
             | 2,000 kcal per day for $7. Likewise 32,000 kCal of rice at
             | Costco is $24 so it gets even cheaper when you buy those
             | 20-40 pound bags. That goes for quinoa, lentils, and all
             | kinds of other staples. Base caloric requirements are
             | really cheap to cover with the basics, and should cost
             | $50-60/mo. The rest can be spent on fresh meat, veggies,
             | and fruit.
             | 
             | Under $200/mo is relatively easy to achieve as long as you
             | know how to cook or can tolerate a repetitive diet.
             | Stretching it to $250-300/mo takes it up a notch and makes
             | it a very balanced and varied diet with whatever fruit and
             | vegetables you want. I only run it up to $300/mo when I buy
             | higher quality meats at Costco and eat an avocado a day.
        
         | boringg wrote:
         | How is Gemini 2.5 Pro v Deep Research? I've found that function
         | on OpenAI quite impressive.
        
           | simonw wrote:
           | They upgraded Gemini Deep Research to use the 2.5 model a few
           | days ago and the stylist shot up - I've seen a bunch of
           | people comparing the new version favorably to OpenAI's, I
           | agree that it's as good and maybe even better now.
        
       | petercooper wrote:
       | I'd rather just be able to paste an API key into the desktop
       | client and PAYG it.
        
         | ChadMoran wrote:
         | I basically do this using self-hosted LibreChat. Can even use
         | MCP with it.
        
       | codingwagie wrote:
       | You can just go on AWS bedrock and use the thinking claude 3.7.
       | You can tune the output size. I doubt you would spend 200 bucks a
       | month in AWS on bedrock even with daily use of claude.
        
         | threeducks wrote:
         | Amazon Bedrock has even worse uptime than Anthropic (see
         | "Uptime" column).
         | 
         | https://openrouter.ai/anthropic/claude-3.7-sonnet:thinking/p...
         | 
         | But I agree with the general point that using the API (with
         | different providers) will have better uptime and that it will
         | almost certainly be cheaper.
        
           | codingwagie wrote:
           | If its not cheaper, just subscribe, but subscribing outright
           | doesnt make sense
        
         | SkyPuncher wrote:
         | > I doubt you would spend 200 bucks a month in AWS on bedrock
         | even with daily use of claude.
         | 
         | I can _easily_ burn $20/hour of credits in Cline/RooCode.
         | Before switching to Cursor, full-time software dev could easily
         | burn $100+/week.
        
         | g42gregory wrote:
         | With intensive coding with claude 3.7, you could easily be at
         | $5/hr - $10/hr. Per token pricing rans the meter very fast.
        
           | joshjob42 wrote:
           | Yeah if you're doing a big coding session up near the limits
           | of context even with prompt caching each reply can be ~80C/
           | or something, so you can definitely blow through $200 on the
           | API if you are using it heavily.
        
       | plandis wrote:
       | Their comparison is not clear at all what you're getting out of a
       | rather large subscription price, IMO.
       | 
       | > Substantially more usage to work with Claude
       | 
       | Compared to pro or free what does this mean? More requests /
       | time? More tokens / time? Something else?
       | 
       | > Scale usage based on specific needs
       | 
       | Are there limits with pro that this is removing or increasing?
       | What kinds of specific needs?
       | 
       | > Higher output limits for better and richer responses and
       | Artifacts
       | 
       | So Claude will do more thinking before responding compared to
       | pro? Is that due to a new variant of 3.7 model?
       | 
       | > Be among the first to try the most advanced Claude capabilities
       | 
       | Okay so you'll get access before people who pay less but not
       | before well known tech people who get early access for testing,
       | I'm assuming?
       | 
       | > Priority access during high traffic periods
       | 
       | If you're paying for pro and they throttle you wtf are you paying
       | them for? This seems like an admission they suck at capacity
       | planning based on their current and predicted user base?
        
       | thomassmith65 wrote:
       | Turned out to be four months, not one                 Leadership
       | at this crop of tech companies is more like followership. Whether
       | it's 'no politics', or sudden layoffs, or 'founder mode', or
       | 'work from home'... one CEO has an idea and three dozen other
       | CEOs unthinkingly adopt it.       Several comments in this thread
       | have used Anthropic's lower pricing as a criticism, but it's
       | probably moot: a month from now Anthropic will release its own
       | $200 model.
       | 
       | https://news.ycombinator.com/item?id=42333969
        
         | zamadatix wrote:
         | This is just pricing the same model to higher limits at a
         | volume discount (which makes the pricing per usage lower, not
         | higher). To actually "release a $200 model" they'd have to make
         | something expensive to run like a ChatGPT 4.5 and then have the
         | only plan which can use it worth a damn be that price. Given
         | how bad 4.5 flopped in terms of performance I doubt many will
         | go that route though, it'd have to be different kinds of
         | services not in the current plan instead.
        
       | aranw wrote:
       | I currently pay for Claude and I still find it best for coding
       | but it's definitely bit unreliable at times and I find it rate
       | limited or responses limited and also unresponsive at times. I
       | wouldn't be upgrading to a $100/$200 packages unless the
       | reliability improved on the $20 offering first
        
         | Maxious wrote:
         | As per https://www.anthropic.com/pricing, the reliability of
         | the $20 offering will be further deteriating to prioritise
         | $100+ package traffic "Priority access during high traffic
         | periods"
        
         | creddit wrote:
         | I had not had almost any issues with reliability until recently
         | and suddenly I'm getting transient UI and server issues on a
         | daily basis. Not happy with it.
         | 
         | Even worse is I bought a yearly subscription from a deal they
         | offered. However, now that reliability will be going down, I'm
         | feeling a bit scammed!
        
       | siva7 wrote:
       | How does this compare to o1 pro for coding? I can't figure out
       | what exactly the max plan offers besides usage limits. Claude
       | also has some interesting novel features but i can't try them out
       | as a European. That's one thing OpenAI does better.
        
         | KTibow wrote:
         | Claude is better at design but isn't as strong a reasoner. I'd
         | expect the Max plan to add a Deep Research like feature too
         | soon.
        
       | m3kw9 wrote:
       | Seems inferior to OpenAIs offering given is same price and
       | capabilities, but OpenAI always have better models first other
       | than code. But code is only useful when using Antvripic API.
        
       | faizshah wrote:
       | Everyday it is looking better and better to self host. I really
       | don't feel like ChatGPT $200/mo is 10x better than $20/mo and
       | after trying QwQ 32B myself I just can't justify paying these AI
       | subscriptions. Gemini Flash is also dirt cheap as an api and Groq
       | is dirt cheap and fast for QwQ and other OSS models.
        
         | parhd9 wrote:
         | api + lobechat has been stellar for me
        
           | ChadMoran wrote:
           | I was using LobeChat religiously until LibreChat added MCP
           | support. Game changer.
        
       | joshstrange wrote:
       | I really like Claude and have the money to spend on something
       | like this if I wanted to but it's not compelling at all.
       | 
       | No new models, no new capabilities, just higher limits. Which I
       | know some people are asking/begging for but that's not been an
       | issue for me. If I needed more I'd probably use the API.
       | 
       | I only continue to pay because some of the features in Claude Web
       | are better than what I've seen elsewhere but their latest web
       | redesign is making me seriously reconsider that stance. It's
       | /bad/, breaks scrolling while it's generating a response, break
       | copy/paste/selection, etc. It's incredibly hostile and I have to
       | assume anyone seriously using Claude internally is using a
       | different client.
        
         | gwd wrote:
         | > If I needed more I'd probably use the API.
         | 
         | The API gets overloaded and you get blocked out as well;
         | yesterday I had to go look up a bunch of bash runes on
         | stackoverflow because Claude's API was busy in the one day a
         | year* I happened to be writing bash scripts. (Maybe I should
         | have tried out Gemini for a change.)
         | 
         | Part of the promise here isn't just usage, but _priority_ : you
         | pay your $200/mo, and (I presume) you never have to worry about
         | being locked out.
         | 
         | * This is an exaggeration but not much of one
        
           | attentive wrote:
           | you can use sonnet3.7 on aws or gcloud and you shouldn't hit
           | any reasonable limits
        
       | annieli442 wrote:
       | A little off topic, but how does o1 and o1-pro compare to 3.7/3.5
       | claude for coding and system design/architecture? I currently
       | have the $20 chatgpt plus and claude pro plans but want to
       | upgrade for a month to do some heavy coding. Thanks!
        
       | spudlyo wrote:
       | If they had something better than OpenAI's Advanced Voice Mode
       | and it came with unlimited access I'd consider it. If Advanced
       | Voice Mode was just 15% better than it is now, I'd find $200 a
       | month to give to OpenAI for unlimited access.
        
         | swyx wrote:
         | as far as i know anthropic has zero voice stuff?
        
           | spudlyo wrote:
           | There is a rumor that they're preparing a voice version of
           | Claude. I assume it's a voice modality rather than simply
           | tts.
        
         | jackstraw14 wrote:
         | Find a way to send them that feedback. I've been doing that
         | more lately and I'm noticing features/bugs I request/report get
         | added pretty quickly. Obviously YMMV but from my experience it
         | seems like they listen to feedback like this.
        
       | ilrwbwrkhv wrote:
       | I have a bridge to sell to all those who are paying these
       | companies $100 / $200 a month. Obviously they aren't hackers, I
       | get that, but even normies should have a better sixth sense?
        
       | jjcm wrote:
       | Aside from the more ambiguous access to newer features first, is
       | there any benefit to this over just leveraging the API?
       | 
       | Going via the API with your own interface seems like it's
       | objectively the better way to handle things like this at high
       | usage, since you'll always have priority, wont have usage limits,
       | and you'll pay as you go.
        
       | AcerbicZero wrote:
       | lol no. Claude Pro gave me like, 20-40 questions every 12 hours
       | before it would throttle, and half of those were mistakes we
       | already corrected that it was trying to re-integrate. Plus Claude
       | is always trying to censor itself. I got suckered in by the
       | "claude is better at code" shit, but wow was that BS.
       | 
       | Compared to ChatGPT, where 20 bucks gets you hundreds of prompts,
       | and if you ask it hard questions it actually answers instead of
       | lecturing about morals, and why you shouldn't be asking these
       | questions in the first place.
       | 
       | Other than using Free Claude to generate resumes to apply to job
       | postings at anthrowhatever, it is by far and away the most
       | useless of the LLMs.
        
       | PeterStuer wrote:
       | Can you even do an agentic approach with a Claude backend as a
       | daily driver these days without something like this $200 plan?
        
       | ashirviskas wrote:
       | That explains why ~24h ago I reached my first rate limit on
       | Claude (3.7 seems to misunderstand stuff a lot and you have to
       | retry constantly which makes you hit those limits faster). I
       | loved 3.5, possibly my favourite model. Might need to shop around
       | soon for something better.
        
       | winterrx wrote:
       | The text for claude.ai/upgrade directs to claude.ai, did AI write
       | this post?
        
       | karn97 wrote:
       | These models are all stupid af. Ask one to make a simple react
       | native page and it will keep getting it wrong. Even O3 mini high
       | is better than 3.7
        
       ___________________________________________________________________
       (page generated 2025-04-09 23:02 UTC)