It probably happens for various reasons. Hence the vagueness. It does sometimes get disconnected due to certain browser actions that would stop it from functioning. Also now I'm sad realizing I have to ditch Comet. I know it has a free tier but Comet was my main browser and cutting the usage down to I think 1/5 by going free won't be worth using that way.
#š¬āgeneral
1 messages Ā· Page 532 of 1
Itās my least used model so Iām not so sure.
I guess I can have it as a backup but I prefer it being my main browser. Makes stuff more accessible
Did not do well on a test search for me
GPT-5 Thinking has gotten better, overall Kimi K2.5 has been the best one so far
Kimi native or in Perplexity?
I havenāt played much with Kimi yet
I haven't used K2.5 enough in Perplexity yet but so far its good as well
But it got a bunch of info wrong on my test search
Interesting, what was your prompt?
I query āNBA games worth watching tonight?ā every day lol
I used it for a bit and it was super useful to have the option to default to a perplexity search in the web bar (wherever the URL is, I'm forgetting what it's called).
I switched back to Brave Browser for the ad-blocking and just set the "home" page to perplexity; not AS nice as Comet was, but better than nothing.
It got the games wrong and didnāt catch injured players
Sorry. To be fair I made it sound like Kimi was better search-wise. I meant their model. I haven't tested enough to know with search and I suspect it wouldn't be as good since Perplexity is used primarily for searching and has been from the start. mb
They have a Research mode, you might want to check that out.
you can change that in comet settings
Yeah I will definitely miss Comet. But hey. Hopefully more agentic browsers drop and/or improve soon
You can disable it here comet://settings/search
Oh well. All this rate limit talk is just pushing me to use Gemini for search more and more
Googleās going to win anyways by sheer will and funds lol
Gemini inside Perplexity or via Google directly?
Google directly
Gemini actively trains on your data and human reviewers look at your conversations though.
Unless you use incognito
How does Gemini compare with "research" functions (like Perplexity Research from two days ago)?
is anyone not able to add files to comet assistant?
Geminiās deep research is intensive
Usually uses hundred or so sources depending on the subject
oh hey you are back
you remember me lmao
And I like the audio overview that generates a podcast on the topic
I left a good memory i see
traumatizing
Gemini is better for searching some things. I feel like the simpler searches are where it shines and can give much better insight in shorter messages. If you think it would be partially relying on training data for the answer, or would be easy to find info (like right there on a wiki page) but needs to be summed up or presented better, go with Gemini. Otherwise (until now) I go with Perplexity research because it digs a bit deeper. But y'know... they just took away 99.89% of Research usage. 
Yeah. My question is... where are we going to go for the tasks we used to assign to Perplexity Research prior to the update?
No idea if this is correct https://www.kimi.com/share/19c2ba9f-32a2-80d0-8000-0000a9d03651
Try Kimi K2.5, the strongest open-source model for visual coding. Explore agent swarm preview for massive tasks. Simplify complex Office work with precision.
A good question. I am gonna try out Kimi but as kt3nma pointed out, it likely is not sufficient
magnificent
you can never unsee this
I will send this to steve jobs
and tim not cooking
you are overthinking it's just a logo
no its men kissing
I just realized why I don't use Perplexity Research, because my go-to has been NotebookLM's Deep Research (yes, powered by Gemini)
apple designers have always been gey
Well, can't wait for the next Grok model to suddenly be AGI and let me just cancel every other AI so my Grok gf can do all my searches for me.
oh! How's that been for you? Is it worth switching to?
I'm constantly using Comet's browsing agent to browse around and do thing for me though
are they created a man pregnent emoji?
š«
I have a theory that man pregnant was a government psyop
by the democrats
and discord/apple were paid money to make it a default emoji
āHow many times do people need to complain about Perplexity removing the microphone button?
The old button let you press it to make the app shut the fuck up and actually listenāinstead of the cognitive AI constantly talking over you and refusing to stop.ā

I like the studio features (i.e. audio overviews and such)
Idk why but I can't find much use for NotebookLM. I like it conceptually. Has some cool features. Never really became part of my arsenal
oh we got a special haircut one
cleaner, better, and also more in detail like a trail log
yeah, it comes down to personal preference.
legit no matter what ppl say, I will pay for perplexity pro
rn I only have Github Copilot Pro, Gemini Pro, and Perplexity Pro
Iām gonna try a notebook LM now because of perplexity recent dumb ass updates with the microphone button
been a perplexity user since like 2-3 years now
it's somewhat similar to Perplexity's Spaces in that you can limit the sources it uses, each has it's strengths and weaknesses and personal preferences.
like aside from perplexity assistant
there arent many
chatgpt atlas + perplexity comet
How do you still have that
I have a folder of all the memes I ever made
with tags
I plan to make an AI agent which will meme like me
Chrome has a gemini sidebar agent thing now I thought I heard
Lmao
no joke
Pair that with nano banana
And turn it into a discord bot
I found old Perplexity pretty bad. Not that it didn't have strengths but in my personal lineup, it never felt like a worthwhile option. Came back quite some time ago and found it to be very worthwhile. Became my main AI. Can't personally justify the 99.89% drop in Research usage though which was most of how I use Perplexity
US only
I dont see a cap on research tho?
Are you on Desktop?
yes perplexity comet
They updated Deep Research #š¢āannouncements and lowered the limits
Yeah that would likely be why. It is 20/month now instead of 600/day. Idk how it applies to Comet though tbh
Was never 600/day
Visit https://www.perplexity.ai/rest/rate-limit/all and it'll show the limits/remaining for your account
Same. Even if it was just 20/day, I could handle that, but 20/month is WAY too low. At that point you might as well just get a Claude Pro subscription and you'll get way more usage out of it than with Perplexity.
Limits sync across devices and browsers
Have you tried out Kimi?
ok lemme see
{"model_specific_limits":{},
"remaining_agentic_research":0,
"remaining_labs":25,"remaining_pro":600,
"remaining_research":0,
Yeah. Research was my main thing I was using Perplexity for.
Not yet. Haven't been inspired to try anything else since Perplexity Research was fantastic enough as it is. How does it compare with its research function?
I can run a query if you want
The limits are lower than what Perplexity offered before this update
well if I ever hit a limit I'll just switch to my Gemini Pro/Github Copilot
where I have higher limits
I usually use github for coding now since I dont wanna use up perplexitys quota, and gemini/perplexity are my research AIs
dont wanna buy more subs
brb 1 hr
Gemini won't let me down. It isn't owned by a greedy corporation... right? 
Anyway gotta work on rebalancing my stack so I know what to use for what. Maybe drop some other subcriptions too if I can find a way to balance it.
yeah, it feels similar to streaming in that if you want all the shows you have to have multiple subscriptions
Depends...

I'm doing research right now and it seems like I might have to switch to a paying Gemini customer instead of perplexity pro user.
Tbh it is worth it just for the various services that it covers.
Why I never cancelled it even when not using Gemini much
I ended up with a Google One AI bundle as before that I had individual subs for smart home and google drive and Nest. Still don't use Gemini as much as I should even though it's now at a higher non-free tier.
How come? What have you been using primarily for your AI?

You get more Google storage (gmail, drive, etc), NotebookLM Plus, Veo Flow, I think Gmail summaries are because of that too idk
Same here.
Perplexity š
yup, NotebookLM has higher limits too
Think that is basically what Plus does right? Idk. Never made good use of my NotebookLM stuff tbh.
yes
Honestly storage limits are why I kept it when not using Gemini. And Veo Flow. The Gemini 3 (I think that's right) dropped so I was back to using Gemini anyway
Oh right I forgot to mention that Gemini also gives Nano Banana Pro. I feel like everyone is aware of that by now but still
if I were to follow the pack and cancel Perplexity Pro then I'd probably switch to Google's stuff but I haven't a reason to cancel yet.
do you not use Research more than 20x/mo?
as said above, I tend to use NotebookLM more often for "research" stuff
I mean if you don't use Research you do not have to cancel Perplexity. I probably would anyway. A 99.89% decrease is just genuinely a slap in the face whether I use it or not
Yeah.
speaking of limits, has it ever been clarified the limits for Comet agent tasks per day/week/month for free versus Pro or Mega?
I guess I'm going to try out Gemini and NotebookLM and stuff to see if it can do what I want it to do as well if not better than Perplexity Research USED TO (which, to be fair, it probably still does, but 20/mo when I'm trying to use it probably 30x/day is just ridiculously low).
curious, what kinds of stuff are you researching?
I know while I was on free Comet, it said upgrade to Pro for 5x usage
seen screenshots of that, but nobody seems to know 5 x (what?)
Yeah idk
It changes month to month. Right now my use case has been job hunting. I have a 2-page tailored resume, along with a 10+ page markdown file that's like a "resume master database" with a TON more information / context about every job I've ever worked, and a specific set of prewritten instructions for HOW to analyze a job when I copy-paste the job title, employer, and job description. It's supposed to parse the job, do research on the company, compare it with my resume, etc, and give me analysis on how likely I am to hear back if I apply with my resume, if I could realistically do that job, and if that job would carry me in the correct direction with my projected career path.
Not coding or anything super technical, but... still, the Research function was perfect for it prior to the update.
if you do use NotebookLM, it lists ths sources alphanumerically so you'll need to rename the re-used stuff like the resume and markdown etc with an underscore or something so it sorts to the top. Then you'll probably need to use deep research to gather sources for each company.
Lmao I think my message got autodeleted for a sarcastic anti-AI joke
where are you from?
Utah, USA
i saw a message pop up and INSTANTLY vanish, I thought you said something mean and changed your mind š
Gemini and Claude Opus both seem to think this is probably the second best option with Perplexity Research now getting throttled down to 1-2% of its former limits.
This is not meant to be insulting but somehow as a non-American your pfp looks exactly what I would expect a man from Utah to be like. Not really a good or bad thing... but like... wouldn't fit California or Texas or... that's all the states right?
pfp?
Profile pic
profile picture
it's even worse, I grew up near Seattle, WA and was transplanted here due to job relocation š
I guess I have no frame of reference for what someone from WA looks like so... that works.
To be fair it is not even a photo so... accuracy probably varies anyway
heh, and that's from before Nano Banana or Midjourney or whatever, I think I had someone on Fiverr make it
From 600 queries per day to 20 queries per day? That's so sad š š
You still have 20/day? Mine is 20/month
I believe it's 20 queries per month now, actually.
99.89% decrease. Which I'm sure is getting annoying that I mention that every 5 mins but... how do they think that is justified?
SMH, I read the #š¢āannouncements and it said they were rolling out the update to Max users ASAP and to Pro users "in the coming days" or something, and figured I wouldn't get the new Deep Research feature until a few days from now. Burned right through my 20 queries so fast.
the announcement did say things are still rolling out, maybe the GUI is showing the same number for "day" and for "month" depending on where you are in the transition.
Yep. Soon enough that 20/day will be 20/month. Thanks Payplexity
Lately, I've been using Gemini for research, probably because it's still 20 queries per day.
My super-broke friend and I used to joke about how much he switches between like 10 different AI tools to get by on free tier. Didn't expect that to apply to paid tiers too after a couple of months.
Now the perplexity for student can use it now or not like a send link to friends still using or not
and... how many free gmail accounts have they created in the past week? š
for perplexity student i finished free trial but didnt start paying, so can i still use the discount and start paying?
Lmao idk. I don't think he has done that. But if there is a free tool out there for vibe coding I can guarantee that I can ask him and get a detailed breakdown of its pros and cons lol
The smart move is to use something like DuckDuckGo's Email Protection feature and generate random aliases that all forward to your inbox.
set up a freebie blog with affiliate advertising and post articles comparing this versus that
Could work
Thinking about that slightly more for a brief moment, I realize... literally couldn't pick a worse target audience. lol
"Let's make an affiliate marketing blog targeting the AI users who can't pay for anything" ...I like these ideas but I have a feeling it wouldn't work out.
ah
Useful blog, for sure. But not a profitable blog.
maybe banner ads instead, only need a click to get paid but it takes a lot of clicks
Banner ads would be better for that idea probably, yeah
bro they have to revert this shit
I love how like a week or two ago they dropped it from 600/day to 25/day and got so much backlash it went back up to 200/day... and then they just double, triple and quadroupled down. It's now 20/month for a lot of people (-99.89% value) or 20/day if you haven't yet been switched to per month.
This feels illegal, but by the sheer power of capitalism, it technically isn't.
even 20/day would be okay, I guess. Use it maybe a bit more strategically, but fine, that's like 600x/mo. But down to 20/mo is ridiculous.
20/day is awful. But it's kinda like seeing someone commit a murder and being like "Hey, that's not okay buddy" and then they kill 100 people and are like "See? It wasn't so bad to kill that one guy" 
"We stopped talking about that first guy I killed. So clearly it wasn't that serious"
fr?
like does the assistant stop responding or what
the limit is basically 10 minutes of automation
thats it
per month
and you cant even prompt it for information after that anymore either
That's per month? On... Pro? Or on Free? 
oh, how do you know how many minutes you used up?
10 mins for pro per month is like crazy
i have pro
Comet Pro users š¤ Research Pro users
[Getting scammed]
i dont know if the limit is prompts or time but i got 10 minutes out of it and then it says You've hit your limit for browser control tasks.
Me, who was both: 
"But it doesn't matter, because we still have our Pro Searches" - Some people for some reason
Perplexity was kinda my go-to recommendation for people to use too. And I encouraged them to use Comet and helped them with stuff. I truly cannot recommend Perplexity to anyone now. Maybe ever. Even if it improves, the fact that they'd silently gut Pro like this leaves a bad taste in my mouth.
Am I dramatic? Probably yes. No, definitely yes. But it's also still disgusting behavior.
very disgusting move from perplexity
Honestly, I wouldn't even be that mad if they were like "Hey, big announcement, very visible on our platform: Starting next month we're dropping from 600/day Research use to 20/month but improving the Research model. Also cutting Comet usage"
I'd still stop using it, but I would be a lot less angry.
Is perplexity comet program still live for some users in Jan 2026 and leads were 10$ per lead and which got to 0$ per lead at the end of Jan so those who earned before that and whos holding period ended and duba dashboard show processed and payout section show pending because payment is awaiting from team is this true will they get paid
I am very inclined to agree with you. I used to recommend Perplexity to people and now I'm hesitant to do so.
@sturdy willow idk any mod can u help me
Yeah I think these companies sometimes forget how heavily word-of-mouth affects their numbers.
We're all looking for the best AI tools, so most of us do tell our AI-using friends what things we think they should use.
is the update just a test or did they say they are not reverting the limits
No, it's actually been updated in their documentation/blogs/etc
I set my name and pic to this last week or so to be dramatic when we dropped to 25/day on Research. I didn't think it would become far more truthful so soon.
I felt silly because it went back up to 200/day but... THAT was the test. Now we're actually being screwed over.
It's showing up to 250 Deep Research/month on their pricing page https://www.perplexity.ai/enterprise/pricing
This is what it shows for me. Keep in mind, this is Enterprise Pro, not Personal Pro. So it's actually higher than what we get.
That is regular Pro
This specific site has been quite weird this past 24 hours
250 Deep Research? at that link I see 50/mo for Pro and 10x Pro for Enterprise which would be 500
Oh, yeah. It is. I got things mixed up.
It was 50, then 250, then 50 again
It's incredibly weird because 12 hours ago, me and a couple other users were seeing 50, now its back to 250 for me.
Best to wait a bit till they finish rolling out the updates I guess
maybe a cached version? I'm seeing this
I honestly think they're intentionally avoiding transparency anyway. If we can have this conversation and still be kinda confused, they're succeeding at confusing most individual users. Which means they can't justify being angry or entirely know what was taken from them.
I've cleared cache, tried in incognito etc
Same person 12 hours ago said it was 250, while me and other users could only see 50
No amount of troubleshooting on the client side sovled it
odd
cause I def want 3 tools to work with
Most things are rolled out in stages.
It really depends on what you need
Copilot is for coding
perplexity for research
chatgpt/gemini for chats/deep thinking convos
Check out ChatGPT Codex
I usually go out on a walk/run and just wanna iterate a bit and find chatgpt/gemini to be good at that
I suppose if I was using Research, going from $20/mo to $40/mo might not be too bad
I have everything included in the github copilot plan
The $20 one?
Codex is surprisingly good in some ways. Other ways, not so much.
idk lemme check
I mean if you are going to buy ChatGPT Pro, you might as well use Codex
given the extremely high rate limits - and GPT-5.2 is a good model
Yeah idk that I could maintain $40/mo but if it meant not getting brutally slashed in terms of Research usage, then I'd at least understand it was an option.
github pro+
I dont code much anymore but its included with the company email I had in my internship so idc
the reason for chatgpt pro is because I have long chats and I just dont wanna run out of context space/get rate limited
Ah, I gotta try out Github Copilot again
and I find perplexity to be ok, but it relies a lot on web searches when im not looking for web searches on everything
Last time I tried, it was significantly worse than Cursor
Even if it ends up as 50/mo Research uses, we can get 20x that by going up to Enterprise Max for $325. So more than 10x what we're already paying. To have 1000/mo. Which is only 33 a day.
which is why perplexity is for research/file gen
I'm trying out Gemini now with Deep Research inside a custom "Gem" (think custom GPT, same thing) but it's taking 5ever.
Perplexity Pro would spit out an analysis for me in approximately 90 seconds flat and have a SOLID response for me.
I can appreciate Google searching a ton of websites and whatnot, but it seems like this route is taking a solid 5 minutes per job and I'm not sure if that's good enough.
Should I be using the "Thinking" model instead of the "Pro" model?
I like the balance of speed and accuracy, ChatGPT just takes way too long
ChatGPT is terrible for searching
ChatGPT Deep Research used to be the best for search
then they just stopped updating it
Gems are outrageous.
I did a comparison the other day and Perplexity with Sonar was better
I believe Gems are similar to PPLX's Shortcuts
Isn't it closer to spaces/custom GPTs?
hybrid-ish I suppose
that's the assumption I was under?
I doubt if even perplexity goes that deep
???
Yep
Dammit. I almost liked David. He agreed with everything I said. But apparently he doesn't get my references. F tier.
The thing is with Gemini Research is it can only produce reports
I can imagine you taking back the friend request and just ripping it in half like "nope sorry"
That's exactly what I typed.
To be fair, when I make a gems joke like that, I never know if I'm referencing the original (Jem and the Holograms theme song) or Taric's line from League of Legends. The first one is somehow less embarrassing so we'll go with that.
I'm unfamiliar with both so I guess you're right </3
which is why I always use gemini for just chats ngl
Okay so Google's "Deep Research" is NOTTTTTTTT what I'm looking for, this is like... it's been TEN MINUTES. For something Perplexity spit out in 90 seconds at most.
(it's at 133 sources right now and STILL doing research)
They train and view on your conversations though :(
i legit only want gemini because its good at my stuff which is all on gdrive
my college gives me files n shit on gdrive and its much easier to just incl. everything into the context and get a detailed breakdown/etc.
Check out some of the OS models like Kimi, Z.Ai, Qwen
with perplexity I always have an irrational fear that maybe perplexity limited me or its not fully completed
Except when it would randomly be like "No, this is serious. I'm gonna do 10 mins of research into whether farts have a color"
all this talk about other models in here and I just now realized this Discord doesn't have an # off-topic channel š
that is exactly what's great about this server
You would get banned in the OAI or Anthropic server
Well this IS off-topic
for mentioning other products
This is the off topic channel
Okay so like right now I copy-pasted a job description for a Supply Chain position at a cosmetics company. I already had Perplexity give me a summary and it was great.
Now I'm waiting and waiting for Gemini's Deep Research and this ain't it chief
It's funny because it says "keep on-topic" while also saying "Talk about anything"
What's even funnier is gemini gave me a mobile app notification 8 minutes ago that my research report is ready but it literally isn't.
it should say "just keep it (kind of) on-topic"
It's mainly for when someone enters an active conversation with something completely different.
I'm going to tinker with Gemini a bit longer (maybe try NOT using "Deep Research" and/or try switching to Thinking instead of "Pro") but if it's not up to par with what I was looking for, I might look into Kimi.
Not comparable - should be comparing ChatGPT Pro to Gemini Ultra
maybe a page reload? sometimes things don't get live-updated when they should
oh no, I double-checked. It just went up to 141 sources now. For some reason instead of giving me the job fit analysis, it's just gathering even more sources.
Deep-er Research
i'm not even joking it said "Synthesizing the Final Analysis Report" and then went up to 154 sources š¤£
just heard a pin drop
Hey fellas šš
What are the best models for business
Hello
hii
Hii
hello everyone i have problem, i have pro version but i cant upload documents
Same here
how much limit in day who knows
Its 50 per day
where can I see how many remain do I have?
for deep research?
For deep research its 20 per month for pro users
oo how many do I have?
Check here
this is my json data?
its 0, i swear I saw 14 yday then i used few like 2-4 at best..now its 0???
this is the ground truth to my account?
nvm i was on a different browser
whats remaining_pro?
i have 600
Search queries
Yea they nerfed it so much but it's powered by claude opus 4.5 now
I remember when it was like 600
do you know if this nerf will be a permanently?
Idk
ugh
Are they doing this to eliminate indians?
Indians got free promo to access PRO
lemme screenshot
almost everyone has Airtel
id say like 70% (ROUGHLY ASSUMPTION)
I got it from magenta
sorry
u indian too?
No
whats magenta? never heard of it before
No
I'm not from usa
Central europe
nice
Here it's called telekom
They did the promotion twice already
When we start talking about the actual limits instead?
why should they? Thinking as their business perspective, they have the full right to do this to customers because customers will stay quiet
is it safe to setup autopay for perplexity?
Alright, so i shouldn't?
if you want them to start and work with us, customers need to stand up. We are their customers and base to their success
but again, would people realize this to stand up?
generally, you should not
they can make u overdrawn
or decide to change the price and make u unable ot pay
etc etc
too many reasons
even if it would never happen to me due ot sufficicent money, I wouldn't do that anyway
better habit
Actually they break multiple of EU regulations, if they want to keep selling that shit in EU they should go with the regulations. They might get reported for shady stuff.
Same with USA
Not surprisingly if India arent going to do anything
lets report them
look at apple for example
they broke regulations
and now they are forcded to change for all
and that change is useful for all countries
maybe EU can be our hero again
u can report them
go for it
In general, Perplexity is a relatively shady company. They do announce new features, but not changes that affect users. If a user decides to buy a service with certain conditions, they should at least clearly announce the changes that affect users. There is no such legal contract in the EU where the text practically forces the user to accept that the service is not transparent. From your statements it is clear that you are a supporter and deny all problems so it's OK.
#šāpro-lounge message
(UCTD) can apply or notš¤ ? By the way, I thought Kacalu13 supported EU consumers in finding ways to file lawsuits.
Hello! In October, I had a Pro subscription for 1 year, but now itās goneāwhat happened? How can I get it back?
You have to add your card details.
Why? The subscription was an educational one, free for a year š Was it canceled?
It's probably because people were able to claim multiple these accounts and sell them. Adding payment details it won't charge you unless your free year has expired.
You "identify" yourself with that card in short.
What should I click here?
https://www.perplexity.ai/account/details
Click "Update billing"
What if I am from Russia?
How did you get the promotion?
Through a friend, they shared the promotion.
I thought nothing would happen when you update your billing, but if they check again, they might suspend you.
My Russian card is not supported:(
On the desktop version, I have the Pro subscription.
Did the tasks page disappear on web or am I tripping?!
Š£ŃŠ°, ŃŃŃŃ
š, как ŠæŃŠ¾Š±Š»ŠµŠ¼Ń мне ŃŠµŃŠøŃŃ
ŠŠ°ŠŗŃŃ
?
ŠŠ° Ń Š¼ŠµŠ½Ń ŠæŃŠ¾ ŠæŃŠøŠ¾ŃŃŠ°Š½Š¾Š²Š»ŠµŠ½Š° ŠæŠ¾Š“ŠæŠøŃŠŗŠ° Šø наГо обновиŃŃ ŠøŠ½ŃŠ¾ŃŠ¼Š°ŃŠøŃ каŃŃŃ, а она ŃŃŃŃŠŗŠ°Ń Šø не ŠæŃинимаеŃ
Ркак ŃŃ Š²Š¾Š¾Š±ŃŠµ ŠæŃŠ¾ полŃŃŠøŠ»
ŠŃли ŃŠ°Š¼ не ŠæŃŠøŠ½ŠøŠ¼Š°ŠµŃ ŃŃŃŃŠŗŠøŠµ каŃŃŃ
ŠŠ° на ŃŠ°Š¹Ńе ŠŗŃпил ŠæŠ¾Š“ŠæŠøŃŠŗŃ ŃŠ°ŠŗŃŃ
Š
Так Š¾Š±ŃŠ°ŃŠøŃŃ Šŗ ŠæŃŠ¾Š“авŃŃ
ŠŠ±ŃŃŠ½Š¾ помогаŃŃ
ŠŃе в окŃŃŠ±Ńе
Š” ŃŠ°Š±Š¾Ńего ŃŃŠ¾Š»Š° Š½Š¾ŃŠ¼ Š²ŃŠµ ŃŠ°Š±Š¾ŃаеŃ, а Ń ŃŠµŠ»ŠµŃона неŃ
ŠŠµŃ ŃŠ¶Šµ ŠæŃŠ¾Š“Š°Š²ŃŠ°
ŠŃ
wait what? I'm a supporter and deny all problems? How?
I supported EU consumers in finding ways to file lawsuits, yes. I was super confused why he attacked me when I was cheering for him
Maybe ENglish isn't his first language
but I hope he can file tho
I'm not part of EU
well, my Gemini said so, lollll.
gave screenshot to it, it said that he might misunderstand what you said blah blah blah
oh!!! ok ok ya he 100% misunderstood me
he prob used chatgpt with a deep research enabled
Oh, my Gemini said that he thinks you tell people to do it, but you donāt actually take any action yourself.
im not from EU. I can't do this. ASk it if i can do it as an indian
ayoo why is gemini attacking me now
what did I even doooo
oHHH
gemin is an AI too
perplexity using gemini
so it took things as personal
OMG
ai has a feeling now
its being biased asf rn
No, it just stated the reason why he would reply you like that.
gotcha
now I see why he felt like I'm attacking them...I AM SAYING in their perspective. i DISGAREEEEE with them 100%
I'm on your side!!! @crimson wyvern
so I felt weird before, but whatever...š¤£
this is actually funny moment
im going back to beating my ai up to learn how to listen
"STT"
Hi guys! I'm back... not that anyone I was talking to earlier is still here. I got muted for 4 hours "automatically" because I sarcastically made a joke that was anti-AI. 
I put "automatically" in quotes because it happened like way after my message got auto-deleted.
My Gemini said that you only explained why the company did that, but you donāt actually agree with them of your messages.
so far, gemini is on my side
good good
ask gemini "is kakcalu13 innocent?"
no no
ask it, does kakcalu13 deserve to get some attack?
Enoughš¤£
why im not able to upload a single image it asking me to upgrade , everytime , i have alread have Perplexity pro package and i dont need max why it happening all sudden now. It was working previously fine i dont know what is happening now
I asked for you, lollll. I combined the answers from the three screenshots.
Same shi šš
thank you gemini! im moving to gemini now
can anybody review kimi deep research, how good is that?
thinking about switching to china model because of the limit issue
what is the best model for deep-search?
good... morning?
you mean deep research?
yes
akhi they already pick the best model for you
for school i use google gemini pro (in perpelxity) and deep research
how i can get pro role here?
Hello
https://www.perplexity.ai/search/kimi-deep-research-review-site-vuuHne2jSmK_4QsSuUQa6g#0
The K2 old version reviews on Youtube.
it is totaly underated then, thx for the info
this is what i need if perplexity still dont fix their limit cap trouble for pro users
@dry kestrel @distant aurora
Gm, how I can reach the team on discord ?
Hey everyone, Iām looking for some realistic expectations on support wait times. I was verified for the Revolut Metal promo and 'Sam' (the AI agent) confirmed my eligibility over a week ago, saying a colleague would finalize the activation. I haven't heard anything since. Has anyone else waited 7+ days after confirmation? Trying to decide if I should keep waiting or if my ticket got lost in the 'high volume' pile.
Guys, Does anyone know why i am not able to see the Deep research option
Itās under the + icon now
FFS why cant I just stay with CLaud? Only because you devs love suckign of Elon doesnt mean Grok is any good. Grok LLM is determined by the IQ of twitter, which is already very low and full of bot spamming.
Now you force model change behind a paywall?
Thank you mate
What do you guys use perplexity for?
I am currently using it for quick searches as it is better to give one sentence answers
I also used to use it for prompt engineering to get better prompts.
Can I get any more suggestions or usecases on how you guys use it?
For context: I am an high school student who will boards in few days, so after this I will have a lot of lot of free time to explore stuff
That's why I am asking you guys
Well just ask Perplexity
Ask for Best Use-case Practices and Brainstorm as Well
Then try those and see what works and what doesn't
I like the api for anything. Create articles, translations, information,...
I was just curious on how the power users use this service, not for only me but for their own usecases
So, that I can learn some system that are really useful, or just to learn some cool stuff
Ahhh i think most of them don't spend time on other platforms except for Work
I hope you find them tho
@young hollow @gloomy cipher @green crystal @dry kestrel https://www.kkinstagram.com/reel/DUTZ6DDCFgR
Ironically, ai might be more forgiving than humans when it comes to recruitment.
So in todayās announcement, it says that Perplexity Deep Research reaches 79.5%
But what is that 79.5%? Reliability?
Try it yourself.
https://www.perplexity.ai/enterprise/pricing
So whatās the actual limit for Pro searches? Is it āUnlimited Pro searchesā OR āUp to 200 per weekā
- New changes.
So oai agent mode tries to order food
It sees delivery isn't available
So it then uses this to hire someone to pick it up for it and deliver to you?
That would be hilarious lmao
I mean, it would be nice if it actually pays.
Free lancing work type shit.
I'm also assuming ai would be better at organizing queues and ratings so even new people get a chance.
A joke. Daylight muggery. Cancelled.
You would get buried under competition on fiver.
LMFAO
add top up credits to my account
@dry kestrel just to be clear, best/sonar is still unlimited, right? Or is it within that 200?
Less than a year ago. Look at the limits now
Weād get 500 per day deep searches, now we get 200 a week pro queries. Howās this even justifiable
Oh looks like a new JOB OPPORTUNITY
Looks Like The Bills are Racking UP...
@dry kestrel crazy if true https://cunnyx.com/i/status/2019188569412124806
Itās comprehensiveness and accuracy in research tasks. You can read more about it in https://arxiv.org/pdf/2601.20975
I guess it only will be for Max

Aravind said theyāre shipping something today, that could be it⦠but Max onlyā¦? Makes sense tbf
Best is unlimited yes @prisma apex
I will agree to that
The Great Ones are Always Busy
They use the tool to ease their work rather than work on their tool to make it more and more better
For now. PPX hardly has a good track record for keeping their limits.
Exactly
As for myself, I was interested in just looking for people who used perplexity to it's maximum potential and made it do stuff that is really good.
Just want to explore on my free time
I see well for now just ask Perplexity or YouTube since you won't always Find The Power Users...
Itās great for anything search related. Youāll probably find some use cases of your own in time
hlo
what is the meaning of "Model Council" english isn't my native language.
Multiple models discuss together and then combine their outputs
imagine them making quests for people like a game
poor man's 5.2 pro and deepthink lol
can someone teach me how to use labs efficiently?
no money
they are broke now
@dry kestrel no cerebras smh
Random, but is there a place for publicly shared pro queries?
Today I got shot in the leg
My perplexity research was 250 two days ago
Now itās 3
š
this is unbelievable
Could be to cut costs, but the value is going down
how do i know my remaining feature credits this month?
Damn
hmm same here, suddenly i dont have credits left, its just the 5th day of the month...
I should have it reset on Jan 24th but its been 11 days and still no reset. Also sudenly research mode went from 200 to 14
Absolutely stupid on their part.
same, just a shame. Even considering that much of the data is fabricated, despite all the deep research you do, there is fabricated data, and now they are reducing it completely; this is a complete disgrace.
fun fact if you write words like email and file it suggest searching with email
exactly
They should atleast offer sonar reasoning pro high. I don't think the current best uses that. I don't understand why they don't offer it. It's pretty competent for research and stuff.
unbelibable, i dont have any query reasearch available, and its just 5 of febrauary.... and i pay 20$ a month
guys if you need a research ill do it for you i am on max its unlimited
Opus 4.6 is Coming
not sonnet 5 is coming
Actually both are Coming
source?
Speculation
I hope they have considered the amount of revenue they will lose with this change
is there a way to complain?
or do thye not check the group
is there admins we can ping? even by email they kept telling me your rates are right and we wont disclose your rates
I think they sometimes do, but the real problem is that AI is expensive, and many companies can face serious cash flow problems if they aren't backed by a strong financial arm like Google or Microsoft. This change will almost certainly reduce their revenue even further, so...
they are very silent now
It shows some here: https://www.perplexity.ai/rest/rate-limit/all
Hi! Im new! Something have change lately with Perplexity, he is unable to read my whole talks, one month before he was able to pinpoint me something that i writed in the same conversation days earlier, if I asked him about it. Now he have atention span of a baby š
Is there some restriction with image generation right now, my country is says is restricted from image generation?
They probably don't know yet until the upgrades to Enterprise and cancellations even out. It might be a wash. š¤·āāļø
they raised $1.5b in 2025, perplexity have money and they good for couple years
No doubt.
Im upset that they keep lowering the deep research queries per month for pro users.
are you sure this works, because i keep getting 600 remaining even when i usse
My understanding is that non-lab and non-research searches resets fairly quickly with most models.
Hey folks š
Iāve been working with LLM apps lately and kept running into the same security footguns:
- LLM calls happening before auth
- user/session data being sent directly to models
- prompt handling that feels risky but easy to miss in review
So I built a small repo-native AI security scanner that looks specifically for LLM-related risks (not traditional SAST).
It runs locally, no SaaS, zero config:
npx secureai-scan scan .
It also generates a proper security-style report (HTML / Markdown / JSON):
npx secureai-scan scan . --output report.html
The report groups issues by risk and highlights what to fix first (auth, PII, etc.), similar to how professional security reports are structured.
This is very early and Iām mostly looking for:
- false positives youād expect
- important AI risks Iām missing
- whether this would be useful in real LLM projects
Happy to share the repo if anyone wants to take a look.
Feedback very welcome š
Hey Perplexity team, Pro user here. First, you took away my ability to generate images, and now you've capped Deep Research to 10 per month. How can it be legal to change the terms of a product from one week to the next?
Comet usage limits are the most confusing thing ever
lol
Im now running the agent on 0 agentic research credits
Why did Perplexity lower the limits on Pro requests and Labs?
they are now too low
20 requests for research and 25 for labs
?
Probably Because The Bills are Racking Up Now...
didn't they release it for some time already
Is perplexity able to get access to the IEEE standards and white papers?
Interestingly, a subscription to Claude also costs $20, but the limits are more generous.
even gemini pro deep research is 20 ones per DAY not month
Probably because Claude didn't Burn as Much Money Perplexity has Perhaps
yeah seed 1.8
Bro
It used to be good, 50 messages in labs and 600 research
seed 1.8 isn't new
i think it was 250 research, 600 in normal search
Well, I don't remember exactly, but I think it was like that.
Pro Search was used to be Unlimited
I remember for sure that the limit in Labs was 50.
best is still unlimited
?
Literally read the msg below that
it's been released for months
The most serious concern is their deceptive strategy: acquiring a large user base through PRO promotions and subsequently altering the terms of service. This is a classic bait-and-switch tactic
Gemini
After gemini launch web assistant does perplexity stand out
I solely used perplexity because of assistant
hi
I can take my bucket off š
No you stay
Infact do you need another bucket?
No thanks
??? What are you talking about
Bro how do you read discord messages
Is my name redacted to you? š
Why does perplexity pro have weekly file limits I haven't even uploaded like 50?
seed1.8
you pinged me with the model card
you can take off the bucket
ok and? dont you know iam autistic
genuinely confused
@fast meteor we're free...
and the cat ears
thats weird i just saw it today it being released but why cant i find the model
nah Iike them
oh nice
seed 1.6 was the one released a while ago
was released in december or something
it's too slow
Is aabb360 still alive
yay
oh hey you are alive
@short narwhal id like to point out that they brought back this menu
ofc just took a long break
Is this research?
Yeah I don't think they ever removed it
welcome back
did it countdown for you??
I swear I've never gotten that
Also for some reason my Perplexity is still on the old UI
The Perplexity gods hate you
yes but the moment i typed anything in or clicked any of the above options the countdown stopped and changed to just a continue button
I think kesku just hates me
so they fixed it
ohhh so I was staring at it too long?? š I guess that's in case if people run it and then go to another tab
So I cleared cache and now I have the learn thing
Yet old UI š
I love max so much so spectacular
still waiting for this
Opus and Sonnet today plsss
model council?
Opus 4.6 is out?
we need it bro
hopium
No
it appears as more of an experiment
than something for regular users
when would you actually use this everyday
gpt is dead in my eyes
people have been asking for this
for a long long long time
Good thing you're not a doctor 
I've been asking for a readback option in perplexity on web for months by now
LOL
OpenAI in general is absolutely dog shit
like this
trueee
why on mobile it's yellow and desktop its blue, i have max too, i would like it to be yellow on both
Idk itās some sort of glitch on mobile it works
just ui inconsistency
Yet Opus 4.5 is the goat
Berd I hope ur a max user
you can force the yellow on pc in comet
claude opus 4.5 is 10x better i sue for everything
how?
k2.5 is the goat for me
Not anymoree
This has been changed
how to get role with max next to my username
Girlll I thought we were comparing expensive flagship models
There's only Pro #š¬āgeneral message
I lost my pink :(
just SOTA models
Why do pro users get a role but us max users donāt ā¹ļø
@short narwhal in your experience whats your everyday model?
i've been leaning kimi k2.5 lately
Gemini 3 Pro
The obvious answer is opus 4.5 thinking
aaaa, does it work for you just fine?? #š¬āgeneral message
yes yes
WHAT
Just buy max
Sorry guys for asking the question here again but -
Hello guys, i have a question. I have the pro subscription for a year, but I'm planning to have enterprise max for a month. Would the enterprise max plan affect my pro in any way after the month is done?
Like would it cancel for the remaining period?
Thank you
just have more money ahh
Itās worth it
Itās an investment at least in your context
Not sure but best to email support@perplexity.ai for this.
nahhh im not maxxing out my limits
okay thank you
I've liked how it uses memory to it's advantage and goes pretty deep with its searches
I did not expect you are a gemini person
It's a love hate relationship
Do you know how ai servers cost to run and models?
yeah but i use pro just fine
You donāt have at least 200$ allocated to put towards something that can further your education?
Quite odd
bro stop flexing your max subscription
Iām trying to get people to buy max pro has been dead
i am on max enterprise it's $325 so chill little player
You people are limiting yourselves
my monthly budget is 50$ bro
you are limiting yourslef by only have max, it's not unlimited like max enterprise which is unlimited
like overall
Max enterprise doesnāt make sense for 1 years
the pro now became 10 and it was 15
1 user
what is happening
i also wonder why do you use enterprise max? Do you run like a company?
@short narwhal there are followup things now in DR
it's for me and 3 employees, startup
yeees been for a good bit
yes this has been here for a month
no the followups
we make ugc style ads, like arcads or makeugc
okay but why would you need enterprise max?
mb I meant the new ui, I need more sleep...
SOC 2 Compliance
and internal sources are nice
whatever that means
He doesnāt have an actual reason but just to impress people on the discord server
i guess im not enterprise
it's an addiction, i use comet powered by claude opus 4.5 to make lisitings, answers emails, and give me summary of what happenend while i was asleep, i give it context so it knows FAQ, and lots of things to make our job easier and more productive
brother you're the one thats trying to shill me the max sub
enterprise slang
It's just an audit thing about data security
On the max sub yes not enterprise
appreciate it
Max makes the most sense logically
also HIPPA compliant
lol this guy is not real
Iām very real
sonnet 4.5, gemini 3 pro
not opus?
doesnt enterprise have the unlimited one?
is enterprise pro vs 2 pro accounts higher limits?
you dont have it thats why not inlcuded it
but i only run opus i dont even use others
the only reason id get the max sub is for the max assistant in comet and if i actually had a job requiring a lot of work at my desk
Not when your monthly budget is 50$
bro let go
yeah thats why i dont get it
On the pricing page they say 2x Pro queries https://www.perplexity.ai/enterprise/pricing
if i had a full on job and expenses id consider it
Howās that even possible though 50$ monthly budget for all your living expenses? You must have a lot external support from your parents
max assistant + unlimited deeep research will make you better at antything, i would be a beginner at someting ask deepresearch and it will teach me anything, i also activate social as source it's very good, get pro enterprise it's $40
enterprise doesnt have the max assistant
bro just let it go
but higher limits
brother i live an average life
nothing special
im still in high school
Ew what is this statement so many grammatical errors
3. Treat community members with respect.
No harassment, hate speech, or malicious behavior. Keep things positive and constructive.
Oh I thought you were in collage of the sorts makes more sense now
i speak 2 languages + english, bro stop being mad at everything it's not that deep, i run a $30k a month business you here flexing a subscription
i agree
Who said I was mad? Seems like youāre quite upset because I pointed out something correct your statement was filled with grammatical errors.
no need to argue bro
Iām not arguing Iām trying to educate this young man
seems like something someone trying to belittle the other guy would say
Are you delusional āyoung manā isnāt an insult.
i am only 21 tho so hes not wrong
i wasted some time here
I doubt it seems like you believe anything people tell you without doing the proper due diligence.
.
Yes youāve wasted valuable time that couldāve been spent paying attention in class isnāt it Thursday why are you even on discord right now. School should be your priority currently young man
My Comet: when accessing the ALT+A shortcut, it's not indexing the current page to respond. What do I need to do? On another PC it works, but not on mine. Is there something that disables this? If so, how do I enable it again?
What happened to the limits of in-deep research?
We're launching Perplexity Model Council for all Perplexity Max users on web. Council Mode lets you delegate to a swarm of frontier reasoning LLMs, where they work async, and a chair LLM synthesizes a more accurate answer considering multiple perspectives.
@short narwhal
woaa
guys why does video generation not work for me despite me being a pro user
of course its for max onlyš
Is this out now?
works like k2.5 agent swarm
also why was recently research and deep thinking removed?
it just released
deep thinking?
the third option labs I think just dont remember the name
renamed to create files and apps
i doubt it
press the + button
did you read the post?
not yet
oh ok thx a lot
should read it before commenting then
does it deploy 100 agents?
ehh isn't that more like it breaking the problem down into different agents? This is just multiple models doing the same stuff and then those things are compared at the end into a final result
what about video generation any idea why it isnt working?
aka agent swarm
havent had a chance to use agent swarm but it did sound interesting
it's amazing
going to do some testing now
though expensive
report back
what is agent swarm?
a lot of releases today
Sonnet 5 where tho
massive and extensive searches like deep diving into companies history or collecting massive amount of financial data
so we got model council, kling 3.0, waiting for opus 4.6 and sonnet 5. also codex model update
which model would be the best for solving olympiad physics and maths problems
OpenClaw sub-agents is the poor mans model council and agent swarm
no single best
havent had a chance to use codex cause ive been using google antigravity and im not really that much of a developer
alr
if I was to choose one then gemini 3 pro
who even is kling
did anthropic always release api access together with the model?
Kling AI
who
oh video gen
and video
zzzz
do you not like my video of skateboarding
of course they rolled it out same day
but deep research advanced is taking its sweet time
Today, we're setting the default for the Chair model that orchestrates individual models to be Opus 4.5. We will keep track of what's best at any given point and update when necessary, look forward to your feedback.
? pls help
they really like their opus 4.5
agent in comet
deep research
and now council
I get why
perplexity is rly amping their stuff up
Knock-off model council is to boot up multiple tabs with different models and then @ each of those in Comet and have one model compare them...
hang on and let me find the kimi agent swarm example i had
too many things going on all at once
is perplexity's max tier equal to the Enterprise Pro tier
time to ask the boss if I can charge perplexity max to the brex
or switch to openai pro
claude is smoking other models
ok model council might not be made for coding
its more for thinking right?
definetly not, its to compare search results or things like tha t
I had this idea of a model council written down on a notepad like 3 months ago
Is this for Advanced Deep Search? Or for something else?
@vague pagoda is model council a reference to pewdiepie
model council
Very cool release!
i remember him saying this a few months ago
meow
he made a model council
this is very epics
I might as well just buy Claud
demn
Bummer it is only for Max, but looks really nice.
Didn't invent it but could have inspired the name or idea even
Claude search isn't great
Is 25 the new cap for labs @vague pagoda š
Wasn't this what we had for lab?
No
literally, i also have claude pro plan and limits are bad, i hit it in 1 hour
Yes
If Model Council would bei on Pro Plan, this would be the argument to get Pro again ... unfortunately only in Max.
Yikes nvm
Pricy
@vague pagoda lemme post gifs
very very costly
Yeah I have a hard time getting good results with their search sometimes
Why is no one talking about this
it is also in openwebUI but it is vs mode i think
you can get claude for $10 instead of $20 with a promo code, search it; there is ton of them
Council of Cl@nkers....is what it should be called 
From a business perspective doesnāt make sense to have it for pro users
CoC
They could like give us a monthly limit
Oh its max only
bro used the c word

We shall pray...
clanka
Clanker
I'm gonna make a chrome extension to bring it for pro users
Itās not even banned why would you censor it
genuinely why need model council
53 steps is insane
for pro users

