💬│general

[potent patrol]

Yeah, as if it would have been a seeling point /s

[dry kestrel]

https://x.com/camerontstow/status/2021322186993107324

Cameron Stow (@camerontstow)

Past threads are now easier to search in Perplexity.

Exact keyword matches are unnecessary. The search will find relevant answers and jump straight to the specific response, even if buried in a follow‑up answer.

▶ Play video

X · Feb 10, 2026, 8:35 PM

finally

[potent patrol]

I deleted all past threats as preparation for the refund.

[quaint ravine]

I thought taht feature came out months ago

[potent patrol]

It's called Deflection.

Deflection is the general term for redirecting focus from the main topic under discussion.

[quaint ravine]

such that I even responded in a pro-feedback thread that that thread can be closed lol

dude, this is general chat, we can chat whatever 😛

[short narwhal]

They added the ability to search contents of threads, but now it also lets you jump to the specific section and follow-up

also what is that font in that video

[dry kestrel]

I wanted that feature for a long time

long long time

[quaint ravine]

ohhh i see

might be more useful to stop some SPaces threads disappearing 😛

that bug has been with 2 of my Spaces for half a year lol

even tried creating completely new Space (with same setup) and same bug eventually happened...quite odd

[short narwhal]

those spaces are behaving like actual space, just floating off

[quaint ravine]

touche lol

tbh, the inline citation bug is the biggest one now

thing is, probably most people don't even care to check

[potent patrol]

Did mesage via that, got an answer from - you guessed it Sam

[quaint ravine]

like for most of their use cases

so won't notice the bug

just to clarify, most of the time, a lot of the inline citations are correct for any one response/thread, but it's like 1 or 2 that are wrong within that response

I have a Space which basically often produces responses that are thousands of tokens long, so there can be quite a few inline citations to check

[dry kestrel]

Did you ask for human support?

[potent patrol]

Did so now. I expect nothing.

[quaint ravine]

also Space used to remember previous Connector settings

despite multiple requests to go back to that habit, nothing has changed, so presumably there must be good reason to keep the current behaviour or reverting to default

[sharp wing]

What?

that for MEE6's comment

this here MEE6's comment

[trim sierra]

Someone knows if theres limits in perplexitys "step by step learning". In free trial?

[opal locust]

why they are restricting preplexity even for pro users

everything is limited now

[potent patrol]

'cause they can

[opal locust]

I just joined the discord to ask you guys if you guys are experiencing the same thing

[potent patrol]

yep.

[opal locust]

I'm glad I got the pro for free

I din't buy this crap

I also have gemini pro, I'll start using that

is it any better ?

[storm iron]

а

[potent patrol]

Please note that we are currently experiencing a very high volume of support tickets. As a result, our response times may be considerably longer than usual. We appreciate your patience as our team works through each request. Rest assured that we will respond to your ticket as soon as possible.

WHAHAHAHAHAHA

I do not accept any delay, as the restictions were also put in place without prior notice.

😉

[modern hazel]

anyone else expieriencing this shitty scam?

i got pro plan, used the automated search, and after 1 search i got this message

[elfin cipher]

This is a scam. I was a Perplexity user from the beginning. Paid for a year-long subscription of 2/5/26, and then they started restricting research queries, even though there is no mention of it. When I asked Sam from customer service, he said there was a monthly limit but could not tell me what it was. I asked him to cancel my subscription, but I was toldI was outside the 72-hour cancellation window. It's really sad because it's a good product. But how can you use a product that has a limit that the seller won't tell you about before you purchase the service, and cannot quantify after you purchase the service? I asked Amex to cancel. They just want to convert Pro users to Max which is $200.

[modern hazel]

you get this message as well ?

[elfin cipher]

Yes

[modern hazel]

thats a shame

we have to make some noise

like, unsubscribe or something or spam this chat

[elfin cipher]

I swear this is what he said when I asked for the limit. "This has recently changed. While we can't share exact numbers, your current Pro plan has usage limits for Research queries. These rate limits may change over time and are essential for maintaining the platform's speed and reliability for everyone. Your Research limits reset automatically on a monthly basis. If you need expanded access, consider upgrading to our Max plan which offers higher limits for power users." I have never heard of anything like it in my life.

[fading ridge]

Wtf is this with the limit overnight, would be great to know who's the flat brain that decided this

[lethal heron]

I experienced this today and their support refused to connect me to an actual person. I cancelled the subscription and reported to BBB. There’s no way they can make a material change to the product we paid for

[timber root]

why this?

[elfin cipher]

I went on Twitter and responded to the CEOs post saying how they fixed the product etc. https://x.com/berz71/status/2021342787497836556

Matt (@berz71)

@AravSrinivas Aravind, I am really disappointed in the company. I have been a loyal user since the beginning. I purchased a year-long subscription on 2/5. (Part 1)

X · Feb 10, 2026, 9:57 PM

[potent patrol]

Look who jost got online

[elfin cipher]

Who

[elfin cipher]

Well you are lucky that the let you cancel.

[elfin cipher]

They just turned of the comments on the CEO twitter thread

[distant aurora]

They'll deliberately keep it slow to reduce demand maybe 🤔

[compact goblet]

Hi @Moderator! I'm from Kazakhstan 🇰🇿. Stuck in an email loop with Sam bot regarding the 1-year Pro promotion. Can anyone help manually?

[distant aurora]

Does it usually show?

[lethal heron]

I still have it for the year. It just won’t renew nor do I get a refund for any portion

[potent patrol]

I'm in touch with apple support. Hope they will pull perplexity out of the appstore

They did this in the past when companies pulled shady tricks with inapp purchases

[fossil shuttle]

New pro limit, it’s way to low

[woven dirge]

hello @signal hamlet could you help me??

[elfin cipher]

Me too and I purchased on 2/5.

[green crystal]

Yeah

[dry kestrel]

It searched for 25 minutes, the final output was useless

It hallucinated a pricing table and there was a lot of filler

[green crystal]

Welp time to wait for gpt-6 lol

For me it's working great atleast

Much more accurate than o3

[modern hazel]

have they fixed it?

[green crystal]

And more Indepth compared to agent

[dry kestrel]

Perplexity did an amazing job https://www.perplexity.ai/apps/42d12472-e5d0-44d9-8176-2ed524a03f12

Way beyond my expectations, and in a timely manner

[radiant pecan]

Just ran into the Deep Research limit, which I've never run into before, and had no indication for how close I was, nor indication for how many I'll have when (if I don't cancel) it resets.

[dry kestrel]

You can see under https://www.perplexity.ai/rest/rate-limit/all

[gloomy cipher]

What mode is this

[radiant pecan]

Does that exist in the settings? I know the Labs limits used to just show on hover, before the changes.

[gloomy cipher]

Research on max?

[chilly bough]

Comet is showing an annoying side bar with a QR code. how to cloise it?

[dry kestrel]

Researcu

[dry kestrel]

People showed screenshots of the limit visible on mobile

[gloomy cipher]

The new opus research?

[dry kestrel]

perplexity

deep research

upgraded

[gloomy cipher]

Yea its opus

The tweeted about it the upgraded version uses opus

[elfin cipher]

Right but if your a pro user it limits you and does not tell you the limit

[radiant pecan]

Ah, I'm on PC. Not seeing it in Settings or on hover.

[dry kestrel]

https://www.perplexity.ai/rest/rate-limit/all

[radiant pecan]

I saw when you posted before, thanks.

[gloomy cipher]

Was qwen not being mentioned here intended?

[dry kestrel]

yes

[gloomy cipher]

@dry kestrel btw have you uses the new qwen max model? It seems to think a lot even on simple prompts it costed me as much as opus 4.6 when i ran a simple request LOL

Actually way more

[dry kestrel]

It's just a mid model

not great but not horrible

[distant aurora]

@green crystal i feel like gpt ruined their deep research cuz I can get better results with extended thinking. What do you think?

[jagged pilot]

Whats now? I have internet connection

[fossil shuttle]

You probably hit the max token limit, it’s starting to forget everything, if it’s a new chat then yeah idek

Which is super low compared to other apps like Gemini and ChatGPT

Probably made it even smaller with the new update

[dry kestrel]

I was talking about ChatGPT

Perplexity used under 5 minutes same prompt and gave me an extremely detailed report

[fossil shuttle]

Oh weird

[distant aurora]

Qwen is overthinking final boss.

The results are good though.

[fossil shuttle]

Bruh you just missed the promo. If you had PayPal you could have gotten 1 year of pro for free

https://tenor.com/view/miro-gif-12838485398045974845

[wind tartan]

no one is gonna pay 200 dollars to use opus when u can pay 20 bucks on cluade to just use it dumb perplixty is cool but im deffintally not gonna pay for free chats you can honestly tell the diffrence between the ai bots and going to the acual ai bots

hope yall work on stuff cause yall are straight up scamming people right now

[fossil shuttle]

Honestly if you just want double limits just go for enterprise pro, it’s $40 you basically get 2X what pro gives

[wind tartan]

limits are no issues at all but i have noticed something it seems to me the models are using the free versions or something they are just so much diffrent compared to going on x and using grok or going to cluade and using it im starting to notice its cool though the research but again can just have gemni run that the costs dont add up or really worth it anymore now that the ais have cuaght up gemini will now run your broswer for you again 10 times cheaper than what perplixtiy wants if yall dont change your pricing and try to keep on comepeteing littarly no point of paying for this service only reason half of us are even using it is the free trail from paypal

[shell tartan]

Any recourse for those of us who paid for 200 deep research queries a day a few weeks ago and are suddenly limited to 50 for the month? Assuming no.

[wind tartan]

ive been testing this for about a month now and pretty sure about this at this point and them limiting opus to the 200 dollar subs like wtf makes no sense gemni cant handle pics and kimi cant when both models can

anyone that is paying for this service is just being silly and wasting money cheaper just to go to the acual ai instead of using this 3rd party cash grab

chrome is also a better browser now than comet now that gemini works with it

no major updates at all just idk was hoping for alot better than all of this not happy and deffntially wont be paying customor

rant over have a good day everyone

[green crystal]

Normal search is more deliberate and direct
Deep research is more shallow and broad and can work for much longer and more sources
It just can't dig as deep and specific
It's also seemingly indeed using sub agents

When you see more spinning animations on the search plan and you open the activity tab to see thoughts
You'll notice it being much faster and sometimes mixing stuff about what it's doing

[green crystal]

Shit now it's back to o3 Dr...

[lapis tide]

I really wonder if i just let my local AI on a 5090 think forever if it could beat any paid models

[clear canopy]

I’ve been feeling pretty discouraged by the current service limits. As a Pro subscriber, it often feels like I’m still on the free plan, which is a bit frustrating. It seems like the recent promotions and free subscriptions given to various schools and companies might be causing some of these issues. While it makes sense that the goal was to build hype, it feels like those of us investing in the service are the ones seeing the quality dip. To be honest, I’ve even started looking for better work opportunities just so I can manage the cost of Perplexity Max.

[green crystal]

if you're aiming for such a premium tier then look at chatgpt pro if you need high precision and instruction following or gemini ultra if you need video, image gen and very large context

[short narwhal]

Not that simple 😭

[clear canopy]

I'm evaluating ChatGPT Pro right now and their Pro version is limited to just less than 10 per month. I have used them as I made comprehensive reports. They said that the Pro agent is a research grade assistant. Sadly I'm stuck with GPT 5.2 Thinking for now. I prefer Perplexity because it gives you access to Claude, GPT, Gemini, and Kimi.

[green crystal]

eh pro tier in chatgpt is the $200 version lol
similar to pro in perplexity is chatgpt plus

limits of deep research are changing now with the rework
im not sure if its still the old limit

agent mode can also function as a sort of dr

[round mortar]

Why can Perplexity do this without any scruples? Are they planning to abscond with the money?

[quaint ravine]

just read the clauses fully perhaps

[wind tartan]

why dont perplixity build their own ai ?

[clear canopy]

I'm using the Pro tier.

[wind tartan]

but i do feel most people will stop using this when the free trails end no one is paying for this crappy service

use cluade chat gpt stinks opus is so beast

agreed brother oh well they either gonna fix this or they will go down hill at this point its seeming they going down hill and will be old news

the free trails were a joke to begin with everything you still cant use

gemini can now pretty much do all auto tasks on the 20 dollar version and cluade is insane and well worth the 20 dollars or 100 if u can do that

i dont belive perplixty is giving the real version capablities either of these programs

[green crystal]

sure? or is it enterprise?

pro tier doesnt show limit resets for models from dropdown

[wind tartan]

i prefer using gemini for deep research and i do belive it works very well

and no real limit problems for me

claude is a good creative writter but amazing coder opus is

of course brother 🙂

[clear canopy]

Business pro

[green crystal]

ah yeah that has limits for pro models

[wind tartan]

i thought chat gpt pro was unlimited

[green crystal]

its mostly unlimited

agent is 400 per month

dr atleast the new one seems "unlimited"

the old one was 250 per month

[fervent matrix]

I dont think any pro rata refund would be possible, same situation I have, not expecting they would care. As T&C also implies Service may subject to change blah blah. But now their reputation sinks to abyss.

I used perplexity partially for work, and now I would definitely not recommend my company to use this, who knows when they will make multiple tier even in business? Or refuse to fulfil their commitments? $200 or more is not a problem for business but IT and Management we have many choices, we don’t like such dodgy practice.

[dry kestrel]

What they showed was not ChatGPT Pro

It was an Enterprise account which is completely different

[lethal heron]

For sure, but there’s definitely a legal argument for frustration of contract. I don’t believe, generally, one party can make a material change to the contract that frustrates the contracts purpose or negates the benefit of the bargain for the other side. There’s a myriad of things that dance the line of civil liability, again, I’d just recommend spamming BBB about it.

[left stream]

holy shit

google hates openai with a passion

they decided to just say screw it and released opus 4.6 for free and expanded usages to all students on antigravity 💀💀💀💀

this is dangerous for openai now crazy asf

LMFAO

[green crystal]

temporarily

[shell tartan]

I think we should en masse request that the old deep research model return at the old quota at least until the end of month? That should cover Perplexity’s side legally as well.

[wind tartan]

as a buisness it would make sense than to go directly to cluade api

same price your spending on here but way better results for your money

[sharp frost]

I feel like perplexity is trying to become a general assistant rather than specalized search/research partner route (since maybe that's where the casual moneys are).

"Best" (non pro) mode is getting better I feel.

[dawn dune]

They were trying to sign a deal with apple to power Siri, explains why Perplexity use to have such a focus on voice chat (and tbf it was the best i’ve tested until then).

[sharp frost]

And samsung I heard

Google integration is hard to beat though

[dawn dune]

Explains why giving away free subscriptions to basically anyone. This model of business works for those who have tons of cash to burn in seek of greater market share

[dawn dune]

Ngl I think Perplexity would have nailed it as Siri integration.
It’s the most vivid and efficient one i’ve seen to this day, but the enshitfication happened and they couldn’t secure the deal

I remember when I first signed up for perplexity getting easily 100+ sources on deep research, now i’ve waisted more then have of my quota and got pretty much most with no even 30 sources

[sharp frost]

So now they have to pivot to the overcrowded casual userbase

[dawn dune]

Yeah, and focusing on Max gives them leverage to push for a higher tier of subscription since everyone has Pro.
The issue is competition has gotten bigger, many LLM’s now have amazing search quality, and Perplexity has not only stagnated but gone backwards compared to when I first signed it

[sharp frost]

Nah I'm disputing your claims about search quality

Not in general use case, but research maybe

[dawn dune]

Compared to what it was before? I think it has definitely stepped up, Perplexity use to be the undisputed best

Now it’s one of many

[sharp frost]

Well I'm only using the big four so far so yea

[dawn dune]

I honestly feel actually bad cuz it was a great service, many times used to get the answer I needed faster and more efficiently.
Now it’s just so much shady practices that it’s just not worth it for the user

[dry kestrel]

Still better than big players like ChatGPT

In terms of both speed and report quality

[dawn dune]

Not sure about that. With 20 per month quota ? No image gen? Coding? Clear business practice ?

[dry kestrel]

Image gen is working fine for me, ChatGPT has similar rate limits for Deep Research

ChatGPT Deep Research (updated today) spent 20 minutes, did 300+ searches, and gave me something completely useless

[dawn dune]

Not really, and so far i’m getting deep research’s where 15 sources are gathered.
And there goes 1 of 20 per month, and if I need assistant about that whose gonna help me ?

[dry kestrel]

while Perplexity Deep Research did it in under 5 minutes, and went above and beyond

[dry kestrel]

Not really as in?

[dawn dune]

As of 20 min ago

And ok issues happen

They probably just updated the backend

Who do I talk to ?

[sharp frost]

Ah yes I had instances where deep research fails early as well with perplexity

[dry kestrel]

One moment

[dawn dune]

There are hundreds of people with either reasonable or non reasonable claims, there’s crazy people of course.
But no one can get it touch

[dry kestrel]

Do you have a prompt example? I'd like to compare

[dawn dune]

Sure but it is in PT-BR

And if that worked for you great! But clearly it’s not working for many out there

And i’ve been here for 6+ months and this discord rn it’s just people clueless why their sub got revoked, or complaining about general service

[dry kestrel]

image gen is working for me right now

[dawn dune]

And there’s nothing to be heard

[dry kestrel]

Sure, I just want to compare results

[dawn dune]

Yeah how many more of that you got?
Compare it as the product as a whole

Realize uma ampla pesquisa em mercado e varejo Brasileiro, e encontre quais sao as melhores opcoes de Headset que competem no mesmo seguimento que o meu atual o headset Logitech G PRO X 2 Lightspeed, e portanto possam ser uma opcao valida de compra ja que o meu logitech quebrou.
Busque por opcoes de ate no maximo 2.000 reais e que esteja em patamar proximo do logitech.
Busco principalmente um fone com alta qualidade sonora para jogos e midias em geral, e deve obrigatoriamente ter conexao wireless 2.4GHz

[dry kestrel]

The majority of users bought promo codes online off a 3rd party service (ToS violation), there have been a few recently that seemed legitimate though primarily from Revolute I think

[dry kestrel]

Enough for the month

[sharp frost]

I joined to complain about the 25/month research limit, otherwise I wouldn't bother 🤣

[dawn dune]

Well great, 20 for me is not enough specially when such poor results are delivered

[dry kestrel]

ChatGPT has a similar limit

[dry kestrel]

Pretty sure image gen is more than 20 a month

at least I've def gone above 20

[sharp frost]

So

[dry kestrel]

Upgraded model = pricier queries

[dawn dune]

Im talking about deep research.
And honestly GPT 5.2 highx which is not even deep research has been better for ages now, and perplexity used to be the goat of search

it takes long, yes it does, but at least I know what i’m getting

Anyways that’s what I think

Hopefully they can get their stuff together, be more transparent of their business practices and step up as they were before, at least in my perspective

[dry kestrel]

I use Sonnet/Kimi K2.5 most of the time and it's enough - aside from testing I don't think I've gone 20+ in a month

[dry kestrel]

It would be great to have clearer documentation regarding these limits

[dawn dune]

Well like I said it if it works for you that’s awesome, but I believe it’s notable that many people do not feel the same. Specially those who saw the path the company took in past few months

[dry kestrel]

I can't read the language but overall it looks right

https://www.perplexity.ai/apps/a9b26176-2f5f-43a4-97a6-c072a9dfb745

[sharp frost]

Can I replace deep research with labs?

Since I never use labs anyway

[dry kestrel]

Short answer: Yes

https://www.perplexity.ai/help-center/en/articles/11144811-perplexity-create-files-and-apps

Create files and apps has access to more advanced tools and can generate files, presentations, images, mini-apps, and other interactive elements that aren't available in the Research mode.

[sharp frost]

Yea for forecasting stuff

Probably even better huh

[dry kestrel]

Yeah

[near hazel]

Can anyone tell me what's happening here?

Like I tried to upload my documents for AI, but it prompted that my "quota" has been fully used up

I'm kinda disappointed with this

Wdym by limiting the docu I can upload per day 💢

Perplexity is becoming worse and worse 😞

[naive sage]

How many uploads do you generally use?

[near hazel]

Like today, I did around 20, I'm a researcher

[naive sage]

Ah OK. I'm sure Perplexity has metrics around how many people do, and they probably averaged it out... but not long ago the plan was updated to be a better fit for 'most' users, and the upload limit was set at 50 per week.

(For the Pro plan)

[near hazel]

If monthly, maybe around 200 uploads, but not more than 400

[near hazel]

Huh that's unfair

[naive sage]

I think Max is around 500 per week, so even Max isn't unlimited.

[dawn dune]

Well I can’t see the actual response but yeah looks good! Maybe not 20 per month good but good. Hopefully in 30 days when my quota gets reset I can get similar result

[near hazel]

They're not trying to cater advanced users

[naive sage]

No, definitely not with the Pro plan at least.

[near hazel]

How can they fulfill researchers or even programmers

damn

Max is so ******* expensive I totally have 0 budget for that

[naive sage]

Are the documents very long? Could you possible get the content and use that in the prompt itself?

[near hazel]

Nah it's just an image including very short paragraph

[naive sage]

There are many other alternatives if you need more file uploads per week. ChatGPT gives about 80 per 3 hours, and Gemini seems to primarily treat it as just more context on a query (so the allowed LLM usage per day might be less - they don't have a set file upload limit)

[glacial anvil]

Merge the images before upload🤣

[near hazel]

Lol if I know there's a limit I'll definitely do so

[near hazel]

Sadly I've paid for one year

[naive sage]

You could upload it to google to get the content of the image, or a local model with image comprehension (i.e. qwen 3 VL 4B is easy to run) to get a description of the image), and then submit to PPLX

[near hazel]

Good idea

Run ocr on another model and then paste it into pplx

[naive sage]

Now it's practically unlimited!

[near hazel]

Lol 👍

[dry kestrel]

I sent it below...

[dry kestrel]

where did you see the 50?

[naive sage]

https://www.perplexity.ai/enterprise/pricing

It's listed as the limit for the Pro plan that it compares to

[dawn dune]

Oh! Let me check

[dry kestrel]

This does not say weekly

[lapis tide]

lol

[naive sage]

It's also listed as the upload_limit in the API at https://www.perplexity.ai/rest/user/settings

[lapis tide]

they removed the old text

https://www.perplexity.ai/pro

now it doesn't say unlimited file uploads unlimited deep research

[dry kestrel]

Does not indicate that it's weekly though

[near hazel]

Ok garbage pplx

[dry kestrel]

This was ChatGPT DR which spent 20 minutes

[naive sage]

Hmm, where did I see weekly... good question. I know it's weekly, but I don't remember how.

Ahh, someone said a support rep told them it was a weekly file upload limit, but didn't disclose how much.

[dry kestrel]

From other users' screenshot, it's showing to be a daily limit

[naive sage]

It also says it's a 'weekly' upload limit in the error message that displays, and here's one of the support rep conversations (there was another one somewhere that I remember): #1468723256277078242 message

So I don't know, all signs point to weekly. Not everything there is a daily limit, b/c Research is listed as 20

[dry kestrel]

Thank you

[naive sage]

I suppose we can clock how long it takes to refresh one quota charge, but I haven't tested it

I wonder if it was intended to be daily, though, b/c 50 weekly seems small from what I've heard about how students use it to upload files, and small in comparison to other mainstream providers... a decent number of them have mentioned it.

[near hazel]

@dry kestrel Very cute parrot!!

[dry kestrel]

Tysm

[naive sage]

How did you make it animated?

[dry kestrel]

its a gif

[naive sage]

Oh wow, I'm surprised more users don't have animated pfps then

[near hazel]

😄

[dry kestrel]

Wait what, my pfp isn't animated

[dawn dune]

It’s good indeed! I would be happy with 20 per day of that
For reference this is day Grok on 4.1 fast thinking https://grok.com/share/bGVnYWN5_3a86d689-f9f5-4fe9-b514-37bf7a010694

[dry kestrel]

I could though

[dawn dune]

150 sources

[naive sage]

Is the icon that displays next to your name not your pfp?

[rigid oracle]

Hey guys. What’s going on with image and video generation? For me not generate any more. What’s change? I’m PRO user.

[dry kestrel]

Interesting

[dry kestrel]

It's like some random user decoration thing

[near hazel]

What is shown

[naive sage]

Oh, also, you don't have to wait a whole month,... deep research recharges at 1 per 1.5 day.

[dry kestrel]

But I could use a GIF as my pfp if I wanted to

[dry kestrel]

rolling window

[rigid oracle]

Region problem

[dry kestrel]

You might have reached the limit

[naive sage]

Oh... is it all from the last query then, and not actually 1.5 per day?

[near hazel]

That's UNCOOL

[dry kestrel]

It refreshes as a rolling window, your last query shouldn't matter

[near hazel]

vpn may help you

[dry kestrel]

This is just from past experience

[rigid oracle]

No, I don’t generate for long time, more than 30 thays

[dry kestrel]

#1140622008086970420

What does it tell you?

[rigid oracle]

https://www.perplexity.ai/search/c6479c07-caaf-4556-bff1-a9b09242bcf6

But the text is Portuguese

Apparently, I will have to contact support by email.

[left stream]

and if it’s temporarily, cybersecurity will come to save me then

I still have perplexity max on the account but lowkey won’t need it

[naive sage]

I don't foresee Antigravity's high opus 4.6 limits lasting long

There's already people who are taking advantage of that... to extremes...

Just wait for the bills to come and Google will take action quick.

[naive sage]

Oh, it looks like the old version of that page wasn't as clear about what that file limit applied to either: #💬│general message

So maybe they'll be tweaking it... perhaps that is a promising sign?

[dry kestrel]

Interesting

[cold sequoia]

Did they limit how many searches I can do in research mode? In the pro subscription?

[left stream]

that’s why adaptation exists and it rewards those who adapt to using the best without caring of the roadblocks as they’re simply temporary detours

[left stream]

100% that’s why adapt in this world full of opportunities that most don’t see!

[modern charm]

Please anyone tell me how to disable this shitty theme! I want to turn the pplx's light colours but can't understand where I got it. I turned off all extensions from the complexity but nothing helped...

[near hazel]

It is a weekly limit

[naive sage]

The real problem, in my opinion, is people using it in ways it wasn't designed to (and which is probably against their terms of use), since people have been using the high Antigravity limits to power things like Moltbot... and, unfortunately, those actions will likely contribute to real financial damages for Google and disadvantages to the Antigravity services for everyone else.

Sure, the people who abuse that may be temporarily 'rewarded' (as seen from one's limited perspective), but they are likely not actions that are free of harm.

[cobalt vale]

what
20 researches per month is a ridiculous downgrade

[queen canopy]

Good Morning

[cold sequoia]

Yea it’s absurd

[cobalt vale]

20 per month and still only 24 sources??? now every competitor has more sources and better limits

[cold sequoia]

Perplexity is basically killing its support base with this

Gg

[left stream]

some use it for evil stuff

thank goodness I majored in cybersecurity so I can learn their methods, apply AI, and then defend

[left stream]

will be interesting to see how it will float in the endgame once it comes close to companies eating each other as the ai race continues

[waxen otter]

In ten days I exhausted my Deep Research searches? I don’t believe it. Something has clearly changed without any notice, and now I’ve been cut off from the very service I subscribed to.
To make matters worse, there was no warning at all. Not even a simple notice like “you have 10 searches left” or something similar.
I contacted support and they told me that for internal reasons they cannot give me an exact number. That’s extremely opaque and feels almost arbitrary.
A real disappointment. I’ve already cancelled my subscription. I’ll have to look for an alternative, because I no longer trust that upgrading won’t just lead to another sudden change.

[naive sage]

This is my biggest concern about the change too. Its potential impact is quite large because Perplexity was so strongly positioned and considered as a challenger for that specific use case - providing people like students and researchers access to high Deep Research quotas so they can get that kind of work done. That was one of its biggest core strengths in almost every analysis, and a key differentiator from its competition. So drastically lowering those limits, even if the new model is better, completely turns that on its head. I've heard people saying that the normal Pro Search should be suitable to doing the same thing, but that's also an area where other providers (i.e. Google) have progressed in the last couple years (and most people who even search on Google nowadays get a basic reply that's pretty good for general questions - and for free (at least currently)). So the change is a real head-scratcher from a product strategy/management perspective. Coupling that with some of the other changes at the same time (i.e. file limits), just exacerbates that problem. It feels to me like that could make this one of the more impactful changes in recent memory (in terms of potential to damage the platform's brand), and so the platform might need a new strategic context/framing in order to really stand out again... almost like a soft marketing event/launch that repositions it with context on the recent changes. Ok, I ranted enough now.

[quaint ravine]

can you share the thread or give me the identical prompt that led to 24 sources with latest Deep Research?

[cobalt vale]

Just tried "Compare the benchmark performance of Doubao Seed 1.8, Xiaomi Mimo V2, and Kimi K2.5 in detail." hoping for more searches, and still only got 31 sources with less than a minute of research

[quaint ravine]

just to double check what do you see when you click the + sign in your PPLX?

does your "Deep research" show that too?

[inner granite]

Interesting

[cobalt vale]

No, but if it's not the updated deep research then it shouldn't move me to the monthly rate limit

[quaint ravine]

interesting

new deep research likely hasn't rolled out to you

[south torrent]

🔒 Message Hidden
-# Discord now requires ID or facial verification in order to see certain messages. Learn more

[royal ingot]

F

[sharp frost]

Same

Too adult

[sacred oxide]

Does anyone know how long does the files generated using Create files and apps stay in the chat?

[plucky sphinx]

Hello everyone! Could you please explain how Perplexity indexes and surfaces websites in its answers?

The issue is that I work for an e-commerce store, and for relevant queries, Perplexity isn't including our brand and website in its responses.

[wide veldt]

hello

[south torrent]

Lol get verification

/j

[sharp frost]

It's not monday but I'm still too lazy

[sharp schooner]

perplexity is getting wrose day by day lol

20 deep reasearches per month, crazy file and photo upload limits
using wrong models for answers

Wtf is that, do they even respond to people?

[rose geode]

lol

[timid frost]

Yeah, so disappointed. I don't mind if they increased the price a little bit, if they maintain the old limit

[wise edge]

WTF? Gemini 3 flash ...

[brittle swift]

Does anyone know if you can transfer the Pro subscription to another account? I accidentally redeemed it to the wrong account

[young hollow]

Probably need to Contact Their Support Email

[brittle swift]

[e-mail] right?

[short narwhal]

Yes

[oak meteor]

Gd luck getting a reply

https://tenor.com/view/waiting-gif-4677175942198662182

[sharp schooner]

only through email but they wont reply

[twin grail]

@dry kestrel

[dire cobalt]

Every single time I try to find something, the first result produced has this label.

It obviously uses sonar in this cases. And Sonar is absolutely terrible model for anything remotely complex. Like for example fetching a video transcript from youtube. The very next moment I press rewrite it with gpt-5.2, it magically becomes available. Interesting, huh?

So virtually after every request I have to stop generation and "rewrite the answer" because I get Sonar "answers". For how much longer this ingenious cost optimisation strategy would last? Can you spare us of it? And can you spare us of Sonar, it is not a model that can do stuff in year 2026.

[twin grail]

nearly 70 minutes reasoning

[green crystal]

With extended thinking?

[twin grail]

ye

[green crystal]

Ahh that's why

They should write the "extended" in large bold letters lmao

[still beacon]

LMAO

[high ferry]

Guys which model is stronger in Perplexity Pro? I wonder if it is Claude Sonnet 4.5 Thinking or GPT-5.2 or anything else.

[distant aurora]

Ai has raised my expectations. I feel like I'm too rude to it sometimes.

[viral remnant]

I think all ai's across the globe are having problems. Someone here above shared 70 minutes of chatgpt thinking so it's having problems. Yesterday on Gemini's discord server even gemini was behaving crazy and now perplexity

[scenic oak]

Hello! I'm new here and I have Perplexity Pro with my mobile plan. What can I use Perplexity for?

[young hollow]

Is just Ai

Not AGI

Still Progressing with Time

[serene sand]

welcome on board!

deep dive ➤ #💬│general message

🤿

[scenic oak]

@serene sand Thanks

[lime steppe]

So im working on an android app and decided to go down a rabbit hole as you do

I asked some about geroge carlin and what he might think of this app crap

https://tenor.com/view/woah-bill-and-ted-bill-and-ted-gif-5444492

[cold sequoia]

Considering they most definitely see the backlash I highly doubt it lol

But yea I’m fs ending my subscription soon this is ridiculous

[lime steppe]

I agree with RoboCarlin here about so called "guardrails"

https://tenor.com/view/bill-and-ted-welcome-to-the-future-gif-9201655

[young hollow]

What is The Reasoning Effort of All Models in Perplexity? @dry kestrel

[peak hatch]

Same was happenign with me few days back. For every first result, it would onnly use sonar, but for rewrite, my chosen model say GPT 5.2 would be available.

[dire cobalt]

yep, and it is "magically" available basically a next instance after the original prompt

[peak hatch]

Exactly

But in my case, it is now fixed for me

could be a bug!

[dire cobalt]

i had it a couple month ago, then it went away, and now back again.
Could be a bug, but also it's a very convenient one for them

[peak hatch]

Indeed!

[lime steppe]

On my way to earning up to multiple pennies a year!

[distant aurora]

@dry kestrel @gloomy cipher @green crystal

They keep popping new models 💀

[peak hatch]

or could be it's their intentional behind the doors bug to cut their costs in one way or the other by lettign the first query to be processed through a cheap model. On a large scale, it could be beneficial

[green crystal]

sadly not heavy enough to force oai to drop 5.3

[young hollow]

Is this Real?

[young hollow]

This Week will be Many Releases

Buckle Up

[distant aurora]

Yes

[young hollow]

Finally

Something Great

Is this Like Kimi K2.5?

[distant aurora]

I haven't seen anything on twitter.

There's this agentic switch though.

[young hollow]

Is it Free?

[distant aurora]

Yes

[young hollow]

Noice

Just like Kimi Agent

Kimi Agent is not Exactly Free tho

[crude jewel]

I'm a Perplexity Pro subscriber and I'm having trouble with the file upload limit.
Yesterday I uploaded quite a few images and received a message that I had reached the daily upload limit.
Today, after uploading only 5 images, I received a message that I had reached the daily upload limit again.

I would like to understand:
Does this limit reset every 24 hours or at a specific time of day?
It hit the limit yesterday around 10 PM here in Brazil.

Could someone help me?

[young hollow]

I really want to See The Benchmarks tbh@distant aurora

Minimax 2.5 is Also Released i think

[distant aurora]

https://cunnyx.com/i/status/2021562855972327636

[young hollow]

Don't trust him

He farms Views i think

[distant aurora]

The official account didn't post anything yet.

[distant aurora]

Huh.

Minimax out too.

[young hollow]

Almost ig

[distant aurora]

Minimax seems to have done a better job here.

https://cunnyx.com/i/status/2021572878236311619

Adam Holter (@AdamHoltererer)

GLM-5 GA Frontend Test:
︀︀Without Skill vs. With Skill

**❤️ 3 👁️ 256 **

CunnyX · Feb 11, 2026, 1:11 PM

[young hollow]

Not Bad

[distant aurora]

https://cunnyx.com/i/status/2021577769977610458

Adam Holter (@AdamHoltererer)

Minimax M2.5 Frontend Test:
︀︀Without Skill vs. With Skill
︀︀I ran out of credits, so these are not the same application. In the base prompt:
︀︀1. One of them is the landing page for the Hivemind
︀︀2. The other is the AI tools page

**❤️ 2 👁️ 46 **

CunnyX · Feb 11, 2026, 1:31 PM

[young hollow]

Another One

[distant aurora]

Lmfao https://cunnyx.com/i/status/2021573089310478773

JB (@JasonBotterill)

@Presidentlin Yeah M2.5 is kinda retarded but maybe its small and cool to run on a mac mini or something. goodnight buddy

**💬 1 ❤️ 3 👁️ 39 **

CunnyX · Feb 11, 2026, 1:12 PM

[dry wraith]

Can we use uncensored prompts in this AI

[brittle swift]

Actually vot a reply within 5 mins

[dry kestrel]

pony alpha maybe

[dry kestrel]

not sure

[dry kestrel]

💀

[low marsh]

Helo

[celest sand]

hi

[short narwhal]

yeah it is

[gloomy cipher]

Glm website?

Bro open router is falling off

Why are new models not being posted

[short narwhal]

Because the weights haven't been released

there's no apis

[gloomy cipher]

Noooo :c

Chinese models are becoming like claude and sam

Qwen 2 image also hasn't been opened yet

[distant aurora]

Yes.

[gloomy cipher]

Imagine if they decide to stop open-sourcing

[vast shard]

How to verify age on discord ?

[dry kestrel]

mhm

[gloomy cipher]

Are you in the UK?

[vast shard]

Nope

[gloomy cipher]

Then there is no age verification

But its gonna be implemented soon

[vast shard]

But I’m seeing messages hidden

[short narwhal]

they're not real 😭

copy and paste it

[gloomy cipher]

Have you blocked anyone?

[vast shard]

Nope

[gloomy cipher]

Sned ss

[vast shard]

#236523452230533121 message

[short narwhal]

that's just text

🔒 Message Hidden
-# Discord now requires ID or facial verification in order to see certain messages. Learn more

[gloomy cipher]

That's not... From pplx

[gloomy cipher]

That means you are from the UK

[vast shard]

Ya this

[gloomy cipher]

Only uk has that

[short narwhal]

bro

copy and paste it

[gloomy cipher]

Where do you live?

[short narwhal]

It is not actually hiding anything

[gloomy cipher]

🔒 Message Hidden
-# Discord now requires ID or facial verification in order to see certain messages. Learn more

[short narwhal]

It's just a silly joke

[gloomy cipher]

matto is a furry 🦄
-# Discord now requires ID or facial verification in order to see certain messages. Learn more

[short narwhal]

https://tenor.com/view/baby-seal-sad-cry-meme-seal-gif-19608082

[vast shard]

🔒 Message hidden
-# Discord now requires ID verification in order to see certain messages.

[glacial anvil]

you missed the 'Learn more'

[vast shard]

Got it now lol

[gloomy cipher]

Sowwy fwien i had tu

[vast shard]

🔒 Message Hidden
-# Discord now requires ID or facial verification in order to see certain messages. Learn more

But this doesn’t say only in UK though https://www.perplexity.ai/page/discord-to-require-face-scans-MWjZI7YXQKKBzEe4th.fWQ

[gloomy cipher]

Berd's hidden folder 🦜
-# Click Here

[gloomy cipher]

Already said it's gonna be implemented soon

It literally says in march

[gloomy cipher]

@short narwhal

[short narwhal]

I'm not clicking that

[vast shard]

So UK got a early roll out then

[short narwhal]

It's either a bird video or a rick roll

[gloomy cipher]

You are missing out

Its all bird

[short narwhal]

I smell it

[gloomy cipher]

Well

It's about berd

And hidden folder

[short narwhal]

🤨

[gloomy cipher]

The berds file

[short narwhal]

are they how redacted

[balmy pebble]

whats the modern 2026 cleancc (circa 2015)

humans vs perplexity

just curious who has better rec

copilot is too dumb

It looks like your message got cut off — “modern cleancc#” feels like the start of a thought, maybe “modern clean C# …” or “modern clean code C# …” or even a question about a clean, modern C# pattern.

Here are the three most likely interpretations so you can tell me which direction you meant:

1. Modern clean C# coding style
   If you meant “What’s modern clean C#?” — that usually refers to:

minimal boilerplate expression‑bodied members
dependency injection records instead of classes for data pattern matching async/await everywhere\nullability enabled

2. Modern Clean Architecture in C#
   If you meant “modern clean C# architecture” — that’s usually: domain‑driven design
   application layer + domain layer + infrastructure layer CQRS MediatR minimal APIs

3. A specific library or tool
   Maybe you meant a package like:

Clean Architecture templates

Roslyn analyzers

If you tell me which one you were aiming for, I’ll dive straight into it.

" cleancc was an open soruce program to uninstall pc software in 2015 - what is the modern cleancc " ask it 100 times get 100 lazy answers -except they do patch bad prompt bugs 247365

"

Modern replacements that match CleanCC’s spirit
These are the closest in philosophy to what CleanCC used to be:

1. Bulk Crap Uninstaller (BCUninstaller)
   Open source Portable version available Deep cleanup of leftover files/registry Can remove Windows Store apps Very close to the “CleanCC vibe,” just more powerful

2. Geek Uninstaller
   Free, tiny, portable Very fast Scans for leftovers after uninstall Not open source, but extremely lightweight

3. HiBit Uninstaller
   Free, surprisingly thorough Startup manager, service manager, etc. Still actively maintained More features than CleanCC ever had, but still not bloated

[dry kestrel]

its a rickroll

[balmy pebble]

puuuurplexity !

https://tenor.com/view/whoa-morerickey-rickey-amazing-superb-gif-11071863372593121250

[sour reef]

Is there a human support here?

[gloomy cipher]

hehe boi

You doubted yourself didn't you

So you checked

[dry kestrel]

there's a nice little button

called copy text

[gloomy cipher]

It's easier to click

[dry kestrel]

that just happens to show the entire message

including the URL

[gloomy cipher]

Just admit your bird ass clicked it

Ik these methods exist

[distant aurora]

https://cunnyx.com/i/status/2021081285767336294

Theo - t3.gg (@theo)

Killing 4o the day before Valentine's Day is honestly beautiful

Quoting Theo - t3.gg (@theo)
︀
Only 4 days left

**💬 133 🔁 60 ❤️ 2.8K 👁️ 203.0K **

CunnyX · Feb 10, 2026, 4:38 AM

Gpt 5 pro getting removed already? While still O3 for deepresearch?

@green crystal

[green crystal]

It's updated but updat vanished for me...

[green crystal]

They're flooding him with death threats lmao

[warm reef]

Anyone switched from Perplexity to an another AI Agent after the new update with limits even for Pro users? xD

[dry kestrel]

Rookie moves

[dry kestrel]

o3 is no longer used for DR

bit annoyed that 4.1 and o4 mini is leaving

[gloomy cipher]

Lmao that's diabolical

No roleplay

[distant aurora]

Looks like they're cleaning it all up.

[dry kestrel]

by keeping 4.5

instead of smaller and faster models like 4.1

[distant aurora]

Does anyone even use 4.5 ?

[dry kestrel]

Don't know

[balmy pebble]

i repeat runfreeze

https://tenor.com/view/spongebob-patrick-lid-jar-open-gif-11485207026862881192

[hot cedar]

Guys, what AI model do you use for analysis and trading?

[civic prawn]

the ultimate question

i’d say opus or gemini

[spark finch]

Opus is really good. Gemini is quite meh.

[distant aurora]

Gpt

[hot cedar]

So which subscription would you recommend for the Pro plan?

[gloomy cipher]

@short narwhal open-source nuke by opus 🙏 https://www.perplexity.ai/page/anthropic-says-claude-opus-4-6-aD3oZbxERcm0fwiwLeG.EQ

[short narwhal]

That's fun

[uneven cipher]

Guys you have any clue why this happening it keeps saying complying is done but isnt generating an answer or a document of it

[solid latch]

8

[modern hazel]

so now it is done that pro plan max automated searches?

or was it a bug?

[serene inlet]

are u f, kidding me, is there one single thing in perplexity that works correctly? They even broke past conversations access today

Does someone know how to fix this?

[ionic hawk]

Hi

[uneven cipher]

Wow crazzyyyy how they nerffed pro

[twilit crown]

Why is Perplexity slowly removing blue?

Now it is mostly black and white

[short narwhal]

Price hike @gloomy cipher @dry kestrel

[gloomy cipher]

BRUH

GLM being more expensive than Kimi

😭

[twilit crown]

Kimi was pretty bad

[gloomy cipher]

?

[short narwhal]

it was not

[twilit crown]

They loop too much

It got into infinite search loops too many times

[gloomy cipher]

if the problem was in pplx, that doesnt make the model bad

[twilit crown]

It loops in other apps too

[gloomy cipher]

kimi in pplx acts weird sometimes

[short narwhal]

Never has for me

[gloomy cipher]

also

in pplx you kinda want it to loop when researching

[twilit crown]

Not research

Just a short question

[gloomy cipher]

example?

[twilit crown]

"What is the latest model from Moonshot AI"

[gloomy cipher]

ok and what is the response?

[twilit crown]

None

It crashed

Because it didn't stop searching

[short narwhal]

Still preferred Kimi-K2.5 over Pony/GLM-5 for everyday use

[twilit crown]

I haven't used GLM-4.7 and GLM-5 yet

How do they perform?

I found GLM-4.6 creating a lot of errors

[gloomy cipher]

same, kimi and claude models have something special to them that feel more human

when talking to them

[short narwhal]

yeess

[gloomy cipher]

like they dont always yap

they are more direct

[short narwhal]

they're nothing special

[arctic sierra]

What's the difference between accessing perplexity through the web interface vs downloading the application for windows?

[twilit crown]

Comet?

Or just the web app

The web app and the website are the same

[gloomy cipher]

which i had the same experience when using pony alpha so hopefully glm is as good

[twilit crown]

Comet is more powerful and can control the browser on its own

[short narwhal]

Kimi and Claude models are so good at language 😩

[dry kestrel]

expected

and also waiting for other providers

[short narwhal]

if only they were to release the weights

And apparently MiniMax 2.5 isn't coming out today (besides in their own apps??)

[gloomy cipher]

dont get high hopes for pplx adopting it

it took them half a year to get kimi

[dry kestrel]

will be released

[arctic sierra]

no, just perplexity

[short narwhal]

Nah I don't gaf about that. I won't even use GLM tbh, just wanna test it out a little

[dry kestrel]

already out on their agent website

and api soon

[short narwhal]

That's what I wanttt, not their own slop chat apps

[arctic sierra]

in the web interface, when I select the user icon in the bottom left, then select "Install apps"

I see two options: One for Perplexity and one for Comet Browser. I have no interest in the comet browser.

[short narwhal]

Give me API so I can plug it in OpenClaw or some other place 🙏

[dry kestrel]

come on it's not slop

[arctic sierra]

Essentially, if I download perplexity (showing as Perplexity Setup 1.5.1.exe) - what's the difference between that and the web?

[gloomy cipher]

this?

why is gemini pro here

[dry kestrel]

yes

[twilit crown]

Same thing

[short narwhal]

I hope that rumored DeepSeek V4 Lite is real and is actually good and even cheaper than V3.2. Would go crazy

[arctic sierra]

ok. Any security issues downloading it vs web?

[twilit crown]

Web would generally be better

[short narwhal]

Difference is you can search anywhere via the omni box, and it's a seperate app if you're into that

[short narwhal]

😭

[twilit crown]

I have it too

[short narwhal]

https://docs.z.ai/guides/llm/glm-5

Some benchmarks and whatnot there

[gloomy cipher]

it just doesnt feel righht

4.7 didnt feel as smart

[twilit crown]

Only Opus 4.5?

They probably had it ready for a while then

[short narwhal]

https://openrouter.ai/z-ai/glm-5 @gloomy cipher

[gloomy cipher]

waiting for tps to be released

[short narwhal]

30-40 🫠

[green crystal]

this here was so predictable lmao

[vocal scroll]

Hey guys, would y'all mind of I share my ai generations here as I have nowhere else to show them off

I've been wanting to show someone what I've been making for the past few weeks with ai tools

[hybrid grotto]

Sure why not, would love to see something unique and innovative

[modern hazel]

anyone else got this message as a pro user?

[hybrid grotto]

Perplexity is literally killing themselves at this point because they can't legally take back the free subscriptions they gave, so they are making it useless, who knows this all might be to sweep in a subscription between pro and ultra and make pro cheaper

[paper spruce]

Hi, my perplexity pro membership which I received through airtel is abruptly paused whereas its supposed to end on 11 september 2026

[short narwhal]

You'll need to add a card to your billing details

[sullen hamlet]

No. Can you understand a company that offers a free service for 12 months suddenly revoking it, asking for verification with the credit card? Before even 5 months have passed. Is the user using the free plan as a student? This makes no sense at all. They give a free period of 1 year and then revoke it for verification, only to charge me for what I used after 1 year? This isn't just happening in the United States, it's happening to me here in Brazil, in São Paulo.

[paper spruce]

Are you sure it isn't a scam? Because the email ids can get spoofed and apparently the sam guy looks like an automated reply generator

[short narwhal]

Go to https://www.perplexity.ai/account/details then

[paper spruce]

Exactly, I'm a student and don't have the credit card eligibility, nor visa, Mastercard or Amex work in my region

[hybrid grotto]

Gemini -Best free
Perplexity Killing itself
Claude too expensive with too less limits

Chatgpt just stably going on, nothing special

[sullen hamlet]

and the copilot?

[naive sage]

Last night I actually got a reply I preferred from ChatGPT more than Gemini... I was like 👀

[hybrid grotto]

Just a wrapper around gpt which ranks gpt which we can use normally via their website as a high end model which would it automatically use for complex calculations

Basically they use their inferior model most of the time and use gpt on complex queries

[sullen hamlet]

Haha, watch out, co-pilot!

[paper spruce]

@hybrid grotto have you got your membership through airtel or bought the membership?

[hybrid grotto]

Airtel free one, I added credit card for it to continue

[paper spruce]

Did it work?

[hybrid grotto]

And then removed it from recurring payment

Yeah

It worked

[paper spruce]

Did any sort of amount was deducted?

[hybrid grotto]

I don't remember, but I can say if it did, it got refunded

They take like 1Rs or smth I think to turn on the recurring payment mechanism on the card

[paper spruce]

Ok, because one of my relative have visa credit card so should I use his? If some amount deducts I'll be held liable

[hybrid grotto]

You can use it but just make sure to cancel the subscription after doing so, check how can you do it for your specific card

[sullen hamlet]

Don't do it that way, it will go wrong.

[hybrid grotto]

And I would suggest you to rather leave it, it's not worth the trouble at this point

Yeah it's not worth the trouble or the risk

[paper spruce]

Okay I see…thank you

[distant aurora]

@young hollow https://cunnyx.com/i/status/2021628386272325956

Zephyr (@zephyr_z9)

Very strong model from GLM
︀︀A bit behind Opus 4.6, but parity with Opus 4.5 at only 700B parameters

**💬 17 🔁 10 ❤️ 248 👁️ 11.8K **

CunnyX · Feb 11, 2026, 4:52 PM

[uncut flare]

Why image and pdf creations always fail? I have pro subscription and I’m in Colombia

[icy parrot]

Whats this app?

[gloomy cipher]

-# MiniMax Agent

[olive gale]

Is it normal that I have restrictions on uploading photos with my Pro subscription?

[tawdry jewel]

I leave this server

I used it only for server tag

Bye, bye!

[quaint ravine]

yes new restrictions for Pro since last few weeks

[olive gale]

If the pro subscription has SUCH limitations, I imagine what the free plan has💀

[lone zephyr]

why am i getting this not even after a month of being a pro sub. i cant upload a image to scan?

[novel arrow]

they reduced limits

[lone zephyr]

ok i saw what others wrote above. before this wasnt an issue

[novel arrow]

i am on max if you need me to deepresearch for you or use opus 4.6 dm me.

[lone zephyr]

no its okay thanks for the offer. i need to use it for multiple submissions with a space i made

think ill just go back to chatgpt

[uncut merlin]

hey dude as a max user do you have unlimited labs?

[naive sage]

It is not unlimited in Max, no - the limit is 250 labs per month on Max, I believe.

I think it's 10x Pro's limit.

[novel arrow]

i never used labs in my life, but things like deepresearch are nearly unlimited i used them and they regenerate in hours

[naive sage]

Think of a webpage or report you want to create and try it out!

[novel arrow]

i tried couple of weeks ago coding and it was so bad, like an ai from 2023

[uncut merlin]

it used to be like that but they lowered the usage from me cause it only says I have 18 usage of deep search left

fells kinda underwhelming

[naive sage]

Try it out in labs, I mean.

[lone zephyr]

i paid the annual plan for pro individual if i upgrade to enterpirse pro annual do i get a partial refund?

[novel arrow]

yeah labs, its definetly using sonar as ai model

give me prompt to try out

[naive sage]

Just something that interests you .. i.e. create a [dynamic webpage OR compelling multimedia PDF report] on [TOPIC that interests you]

Can expand the prompt from another model first so it'll do better... or fill it in with things that you are specifically wanting to see... but if you've really never used labs it might be worth a shot.

Just to see what it can do.

Maybe some 'samples' in the app would be helpful to that end, some other AI providers show samples for tools like that so people can get inspired and/or figure out how to prompt better.

[novel arrow]

i tried it and wondering what sdifference between deep research besides the fact that it generates chart and stuff

[naive sage]

I think it has more agentic tooling/capabilities for generative assets and different tasks (mostly code-related)

And can generate more than one file at a time (I think)

[lone zephyr]

what do u guys use max for?

[novel arrow]

work

[meager rover]

Why am I just getting this? Hasn’t 4.6 been available?

[lone zephyr]

cool are you a lawyer?

[valid charm]

https://windsurf.com/leaderboard

New benchmark from Windsurf

[novel arrow]

ecommerce

[minor shard]

Wen sonnet 5

I was promised sonnet 5 😞

Claude is really killing it for coders.

Kinda crazy

[lone zephyr]

ohh nice

[lone zephyr]

claude is really good with reasoning

[carmine ember]

which model do you recommend the most?

[gloomy cipher]

the bars are so thick it feels like its made by Lab 😭

[jovial plinth]

im cooking

[lone zephyr]

depends what ur work is for most of mine is using claude 4.5 with thinking enables

working with maths

[jovial verge]

Is this lame - they can’t say when Pro sub deep research credits end? My plan sub got renewed yesterday and today I can’t use Research?

Tried logout from all sessions and restarted. And yet the same result

[quaint ravine]

you can check Research limits here - search for "remaining_research"
https://www.perplexity.ai/rest/rate-limit/all

you gain back 1 Research every 24-36 hours or so (based on personal observation, and also others posting here)

[quaint ravine]

one way to look at the new Deep Research may be that it allows Pro users to use Claude Opus

for 20 bucks a month

[crude spear]

what is this browser called stack?

[short narwhal]

gemini-3.1-pro-preview

https://cdn.discordapp.com/attachments/1237372301243191327/1274608969989754911/catBOOMM.gif

No clue what kind of tests you're looking at

[celest drift]

I’m flabbergasted at how my 240$ a year pro account is massively limiting my Deep research searches. Even more infuriating is that this change was made without notifying paying customers. I really hope the team gets it together or I will leave the paid account and go to another provider.

Has anyone gotten clear answers about the new limits? Or does anyone have a contact at the company.

What a disappointing customer experience

[short narwhal]

It's more limited now due to it being much more powerful #📢│announcements message

Look more at the contents instead of the word count

[celest drift]

No notice is insane. I’m hoping for a human response or a change or I’m canceling.

[short narwhal]

https://www.bloomberg.com/news/articles/2026-02-11/apple-s-ios-26-4-siri-update-runs-into-snags-in-internal-testing-ios-26-5-27 @dry kestrel I'm going crazyyy, this is never getting released

[dry kestrel]

hahahah just saw Mark Gurmans post

[quaint ravine]

time is short for everyone 😛 If you only need to read 1000 words to capture everything, why bother with 5000 word filler 😛

it depends what one is after of course, but i think in general, being concise and accurate (while not missing any major points) is better, which is what appealed to me with PPLX's style from the start

I started my "serious AI journey" with Gemini, so I know what people mean about it tho lol

[distant aurora]

@green crystal people will read this chat and blame ai for someone committing suicide, lol.

https://chat.deepseek.com/share/ojlahl14wabb7ig33z

I on the other hand like how freely and honestly it speaks.

Gpt would probably get stuck in the safety bs loop.

[dawn dune]

@dry kestrel I give up

Idk Imma try to erase all memories

[wise nova]

Already available at chat.z.ai

[jovial verge]

All I see is zeros for remaining although the billing renewed yesterday only!

[jovial verge]

I’m facing the same issue. The monthly bill date was yesterday and then today with hardly any use, I have zero remaining Resarch queries. Just subscribed to GPT also now? Not sure if the pro sub here makes any sense for me with this limitation. Maybe Perplexity ran out of compute resources and throttling a lot

[celest drift]

It’s literally insane. I think I might try other options. Super disappointing. I’ve been paying for a year.

[slim wind]

i wish there was a way to go back to the previous model with the higher limit

kinda scummy not mentioning anything in announcement about lowering the limits

but ya know, if they want to lose customers faster

[hollow garden]

[slim wind]

good on them

they are 100% doing a great job at that

[true ginkgo]

Why can't i find my damn referral codes??!!! Did they take that from us too??!!!

[shrewd wedge]

woww there’s referral codes? 😭

[jovial mantle]

GLM-5 won't stay king for long I'm afraid, deepseek is in the process of updating

[novel arrow]

Perplexity reduced limits for max users 😂 😂😂😂😂😂😂

Any other max user to confirm

[pine talon]

[slim wind]

Image uploads have also been decreased?

Bro wtf is going on. Why has none of these changes of limits been announced

Guys cmon. Dropping the ball rn

[distant aurora]

You want opus free?

[wise nova]

Yeah,

[distant aurora]

Go check arena.ai

[wise nova]

I know, using it for two days. But web ui has bugs and mostly stopped unexpectedly, asking for generate again

[distant aurora]

They collect and own all the data and chats btw so be careful about what you share.

[distant aurora]

I faced this issue too with gpt 5 search.

@gloomy cipher @green crystal bro knows Epstein

[gloomy cipher]

Yea websearch bro

[wise nova]

The second screenshot shows the generation process which lasts 15h

[distant aurora]

Why the restriction though.

[gloomy cipher]

It's not restriction to him

It's probably because the information is nsfw

[wise nova]

But the results are amazing, Opus thinking understands at a glance, gives information to the point and sensibly

[naive sage]

Why's everyone just focus on his controversial aspects. They could at least list something rather than nothing... word it in another way, provide a more general background/statement...

[gloomy cipher]

Qwen has a guardrail model behind its chat

[distant aurora]

Eprey Jeffstein

[gloomy cipher]

It itself is uncensored

But a smaller model decides if things are nsfw or not

[slim wind]

??

[gloomy cipher]

So like

[naive sage]

Like that's not a good summary of the person.

[distant aurora]

I just wanted to invert his name.

Epprey Jefstein

[gloomy cipher]

A smaller model runs the prompts to check if there is anything that it shouldn't reply to. Because qwen os inherently uncensored

[slim wind]

Pretty sure that’s what he is known for

[gloomy cipher]

That's why it sometimes can be bypassed or fail

[slim wind]

Unless you have something good to say about him

[distant aurora]

Btw

[naive sage]

It'd be nice if AI just didn't regurgitate the most common complaints about a thing and actually get to the important bits.

[gloomy cipher]

Bro say it, i wana leave discord

You are making me wait

Peasant

https://tenor.com/view/angry-emoji-face-mad-gif-27250868

[slim wind]

true but this is jeffery epstein. should it not say that he was a pedo?

what would you want the ai to say

[gloomy cipher]

Bro delete this

Why are you sending these things in a server

[naive sage]

Something more toned down is better than an empty response.

[wise nova]

Claude Opus 4.6 Thinking: generating 9 page docx document, 74% of a session limit, 7% of weekly

[slim wind]

ah ok

[distant aurora]

Spy

[wise nova]

How to upload and attach files in chat?

[distant aurora]

Only models with this icon support fine uploads.

[wise nova]

Only pdf allowed, no txt, no md

[distant aurora]

Yeah.

[cerulean sedge]

Might be a stupid question, but i notice that on my laptop it doesn't tell me how many DR questions i have left for the month but on my phone it does

Does this mean that on my laptop I have unlimited? Or do I have the same but it just doesn't tell me

[slim wind]

pretty sure pro has like 20

or 3

idfk atp

all i know it went from "high limits" to very low limits

[short narwhal]

It’s the same limit. You should be able to see it by hovering over Deep Research. Or just see it at https://www.perplexity.ai/rest/rate-limit/all

[slim geyser]

does PRO still provide video generation? I have a new account haven't used any credits and its prompting me to upgrade to MAX. According to Perplexity Support themselves video generation should be available for PRO users in Canada however I'm still unable to get it to work.

https://www.perplexity.ai/help-center/en/articles/11985060-generating-videos-with-perplexity

[cerulean sedge]

Okay thank you

Yeah on my laptop it doesn’t say

But does on my phone

[delicate dawn]

PRO USER GENERAL FEEDBACK PLEASE:

Can you confirm that, before the last couple of weeks, limits on features like Deep Search and Browser Control were far higher than they are now?

Personally, I used Deep Search for work between 6 and 12 times a day before these unannounced limits came into effect. Was this your experience too?

[green crystal]

@dry kestrel do you still see the new oai Dr?

[dusk skiff]

Is this happening to anyone? I’m a pro member and now I can’t even use the other models. Also they deleted my history of threads like 6 months or more 😭

[cerulean sedge]

Yeah, I used it often for school up until now and never even knew that there was a limit. Now it's very limited and especially with the way that you don't get all 20 uses in one go i'm concerned about whether it's enough

Now i'm considering whether I even want to continue my subscription honestly

However I will acknowlede that i'm in law school, and i've found that Perplexity was the best model for helping me with that

[slim wind]

may i ask how you use perplexity for law school?

just curious

[dry kestrel]

Yes

[green crystal]

Weird for me it's now entirely gone as if I never got the update

[cerulean sedge]

I find it's best for searching up cases and doing case briefs and stuff like that and citations

I note for example that chatgpt may tend to make up cases or facts within a case that perplexity wasn't

perplexity imo is just more research and fact based

I used to use perplexity to help with the research and note making aspect and another service for helping to put it in essay form (in the context of essay)

[wind tartan]

do the mods even read anything in here lol

so many un happy people lol

[cerulean sedge]

Well tbf i'm sure they all knew people would be unhappy with the change

Who would be okay with seeming unlimited DR to suddenly 20

I'm sure the reaction isn't anything they didn't account for

[wind tartan]

time to move on to chrome google ai now does everything perlixty does and cluade is a better as well buying the script over paying perplixty these guys after all the free trails end will have no custmers at this rate

i honestly dont think they care it was just a quick cash grab for them

[slim wind]

might be me but weird

nvm

its fixed

[brazen pine]

I’m looking into Kagi

[broken plinth]

why i am not able to use deep research even though i have pro account is perplexiy no more working :

[shrewd wedge]

[brazen pine]

They cut the deep search from 400 to 20

[broken plinth]

20 deep search pr month ?

[cerulean sedge]

Yes

And you also don't get your 20 in one go, it refills 1 every 1.5 days

with a max of 20

[naive sage]

So 20 is the optimum use case, if you don't query once every time it resets, you likely will only get some number less than 20.

[broken plinth]

It said I have already finished my monthly deep searches, so I am no longer able to do any more.

[cerulean sedge]

Even this is annoying because i would definitely use all 20 for school, but i'd want to let it build up to the full 20 before using

So I can just do all my notes in one go

[junior yarrow]

They made changes without informing us. There’s a petition going around everyone with pro impacted by these changes should sign

[cerulean sedge]

my main gripe really isn't the 20 use limit (although i believe that's a very little amount for a paying customer), i just think the restoring 1 every day and a half as opposed to just refilling every month is very impractical

[broken plinth]

they are indian startup whatelse can be expected from them sue them

[brazen pine]

they dont care about no friggin petition. better off searching for an alternative. Might just do claude directly

sue for what? Its their product they can do what they want

[junior yarrow]

It’s a San Francisco business, they are held to a higher standard doing business in California.

[safe fractal]

Where’s the petition?

[brazen pine]

too many services out there doing the same thing. im moving on

[shrewd wedge]

reminds me of tv streaming, so many services u end up paying so much

[dry kestrel]

any VPNs?

[dry kestrel]

They are a U.S company

[slim wind]

How would one sign?

[naive sage]

That's why the tech bros usually meet up and make changes like this in tandem. 🙂

<tinfoil hat>

[dim blade]

Bummed that Deep Research got nerfed

[obtuse coral]

Okay so im having this issue where perplexity api is giving me utc time zone but i need est time zone anyway of adjusting this so my model has correct time and date.

[hidden sentinel]

Solution: switch AI's cuz perplexity is a scam

[obtuse coral]

hyeah bro ive wsted money left and right for a pro account api keys busniess account and still i cant get the correct data set and being locked out where i cant even send files

[marsh marsh]

Any discount code that i can use to get pro subscriber?

[obtuse coral]

i think im in the hole now 200 ish dollars

[shrewd wedge]

ouch

[hidden sentinel]

don't please

do not buy pro

its a scam like deadass

[obtuse coral]

so fucking annoyed right now

[obtuse coral]

dont waste your time legit

[meager rover]

yup pro is absurd

like spend the same amount of money on ChatGPT Claude or Gemini

[obtuse coral]

once you get the pro account you run into laggs and limited data

[dim blade]

Paypal Business was giving out a free year sub for pro a few months ago. Not sure if still available

[obtuse coral]

dont waste your time after this month im switching my model brain( currently perplexity) to something else with stronger reasoning and shit

[shrewd wedge]

claude ftw sorry not sorry

[marsh marsh]

I want to use their voice speak tool after research. Any AI tool has this function?

[obtuse coral]

not worth it at al even if its free

[dim blade]

I mean, free is free, but the nerfing is getting out of hand for sure

[obtuse coral]

free isnt free if you cant get the work done that you need to if they keep asking for funds

at this point you figured a staff member would see our concerns and address it rather then doing a token spike throughout the day and losk you out from adding files. legit went through $20 of api usage but i barley touched the model and i had to reload the account again. so they are defintialy increasing the token usage and making it seem like it you whose using all the token in day

please explain this to me

[shrewd wedge]

google has 1 year of free ai pro for students

[obtuse coral]

how tf did i use more today then i did when i started buliding the model earlier last week

[shrewd wedge]

antigravity got opus 4.6

[obtuse coral]

that ill use

[shrewd wedge]

gemini 3 is so shit for coding tho imo

4.6 and 5.3 >>>>

[brazen pine]

i was about to say that

ima just hit claude and keep it pushin'

[obtuse coral]

hmm yeah that not good so which would be ideal to use for coding

[shrewd wedge]

claude so expensive tho , codex is pretty neat

[obtuse coral]

basically i need a new brain for the mode( better reasoning then perplexity api)

[brazen pine]

claude is $20 as well

[shrewd wedge]

claude u hit limit so quick 😭

[obtuse coral]

yeah i nned one where i dont get limited right away

[shrewd wedge]

what are u building

[obtuse coral]

im trying to build my sport betting model so alot of testing is required

[shrewd wedge]

codex

use grok 4 fast for testing i guess

so cheap

[obtuse coral]

alos im not code saavy so does it have ai assitance imbeed

embedded*

[shrewd wedge]

i read “in bed” first 😭

[obtuse coral]

lmfao that were im about to be in a few minutes lol

[shrewd wedge]

i already am 😴

[obtuse coral]

so this is what the model is looking like in terms of profit and stuff , so yeah i need to test the hell out of it so i can get it closer to 90% for nba

[shrewd wedge]

sweet

[obtuse coral]

here the current status update i just ran

[void coral]

Hi there, I have pro subscription and now I am having trouble with pro research. Is there AI substitute for perplexity AI?

[twilit lance]

I asked claude/sonnet about what the upload limits were per chat per day, didnt say ANYTHING about perplexity other than im looking for an alternative to it and... uh... Sonnet might have its own opinions about whats going on rn lol

[obtuse coral]

Amazing

[twilit lance]

Had to share

I have never used it before this chat too so it’s not like it bled over from a previous convo

[obtuse coral]

Claude is already talking shit about perplexity makes it even more comical

[naive sage]

It probably just inferred the reaction from the user's problem statement in the prompt, not from articles.

[twilit lance]

My brain is creating an image of it being offended by the way perplexities internal commands work when you use sonnet through it

[warm lance]

Hi

[twilit lance]

I asked for an alternative to perplexity, I wasn’t baking a negative tone into it

[naive sage]

I haven't seen anyone complain on youtube yet, it's still just shills saying how revolutionary the new Deep Research model is.. it changes everything!

Youtube just gets worse and worse every day it seems

[twilit lance]

To be fair my issue isn’t the quality of the tool it’s the limits

[obtuse coral]

Yeah exactly

[twilit lance]

Though the response quality consistency has been a lil… interesting lately

[obtuse coral]

And the hidden token usage spikes they hit me with tonight

[obtuse coral]

You mean completely off and retasking the same script loops where it keep having you redo something that has already been created

[twilit lance]

Yeah or it dies mid answer and then you resent the prompt and it suddenly can’t see any of the previous chat

It… happened before and now it’s a daily occurrence

[obtuse coral]

Yup which make the whole spaces thing useless

[twilit lance]

I wish it was better able to read space docs

[rocky yarrow]

Guys did they fix Perplexity yet? It's still bugged for me and won't do deep research

After like 2 days if working

[twilit lance]

Do you know about the fact that its not unlimited anymore

it may not be a bug

[rocky yarrow]

What

[twilit lance]

.... yeah...

[rocky yarrow]

Where's the announcement?

[twilit lance]

im so sorry to be the one to tell you this

There are monthly limits and they are...

[rocky yarrow]

I paid for a year long subscription 6 months ago for unlimited what are you even saying rn lol

[twilit lance]

how do I put this

not great

I found out today myself

[rocky yarrow]

I went on vacation and get back to work yesterday and now I can't even use pplx anymore? Lol

I don't believe you guys

[naive sage]

Most of the limits are posted here (but not all): https://www.perplexity.ai/enterprise/pricing

It compares the Pro plan limits to Enterprise

[rocky yarrow]

Where can I talk to the perplexity team? This must be a bug

[twilit lance]

Im right there with you in wishing it were