#💬│general
1 messages · Page 528 of 1
Hi guys I am facing an issue, can I ask here?
Or is some other chat more appropriate
This is my first time here and I don't use discord much, sorry 😅
It is difficult to say unless we know some details around the issue. Maybe... maybe not... it might be related to something else that people have been asking (if you have tried searching the server for some keywords related to it). If it is a billing issue, then we can't help, naturally. There's a lot of things that 'an issue' could fall under... this reminds me of one of Berd's favorite sayings... "don't ask to ask"... 🙂
Is it giving a message about your region being restricted?
Yes
I've had that before. What causes this?
Hmmm
If so, I would try a different image generation model and also recommend contacting support and asking if it's accurate that your region is subject to restrictions. I don't think it's been published anywhere, but it's clear that several regions were restricted a few weeks ago. I've speculated it was related to the xAI controversies in the media and the EU lawsuit... so either a provider and/or perplexity themselves made a deliberate decision not to provide image generation in several places. I had heard that Germany was one that was affected, some other places in the EU/UK, and someone said they were from Czech.
When that kind of thing happens, communication channels need to be better... otherwise we get this as the result, lots of people coming in the Discord and asking about it.
The community gets to try to figure out what changes were implemented from the complaints, it's not ideal.
Hello all
.
When are we getting Kimi 2.5 on pplx?
Was the free one-year Pro subscription canceled? I suddenly see a price of $4.99.
Is there a place to see AI prompt for images?
hello guys… I need help, i got an email where Perplexity asked me to add my credit card (I already put it to have the Pro subscrition) and when I try to add it… it says that I can’t add it??? it says I will lose my subscription, can someone help me? This is annoying me..
Image generation is still broken after a month?
I have received the same email. Is it legit or some sort of scam email that is being sent to the users?
i was thinking about this too.. so if someone know pls help
lifehack
cancel pro subscription and system will offer to you discount 50% for 3 months (10$ instead)
Is it planned to implement Kimi 2.5 agentic in the list of the llm?
hi, tell me
ehmmm no, i won’t click links, you don’t seem a moderator.
@signal hamlet
nice try diddy
Watch out guys, this is a scammer
@signal hamlet he tried to add me as friend, hope this helps
Thanks 🙂
Hello all
So funny since on my old perplexity account, Research Pro got down to 200 and sometimes even less queries a day, i cancelled and started a new account and i got 600 again 🤣
So what’s going on with perplexity recently?
Some limits on pro users. It seems to be okay for now.. Not to glaze it though
A while back i had Perplexity pro and i used it and then i stoped for a few week but now the pro version is gone and i need to add my card info to use my pro version can you tell me why it may be
Hello. My 1-year Pro trial comes from my mobile carrier. I got this email too, so I decided to play it safe and use a Revolut virtual card with a £1 limit(I use it for all subscriptions like this)
Perplexity site said the charge today would be £0, but it still tried to charge me £73, and then there were two more £0 attempts. Same thing happened both times I tried. It's sketchy to me🤷♂️
really?? Omg.. ty for answering dear
Welcome. I advise you to be careful and also set a limit before adding a card. Just in case.
same, but i had no problems it was smooth and legit, it didnt tried to do anything
Thank you guys❤️❤️
So it didn't try to charge you anything?
no, nothing
i didnt even get a notification from my bank, and i still have all my money
I don't know why it was different in my case
and i checked my perplexity account and it was there
might be regional issue ig
It's an anti-fraud measure. You won't be charged until the end of your billing period, and adding it will continue your subscription
Who helps with Customer Service?
ow alr
nice pfp! ^^
hi guys
Do you know if pplx actually limits models thinking/tokens limit? Did anyone actually test that, and can I do that, maybe there are tools?
Thx!! And you too!!
❤️❤️
whisky says thanksss ahah
It tried to charge me £73. And my trial ends in October. Maybe I'll try later with other card.
i used gemini 3 pro a hella too mucha and i got a message that i hit the limits, but it takes a lot of tokens, from my experience
I'm using 3.0 as well, what kind of a limit was that?
Like daily pro requests?
Also my question was mainly related to claude sonnet, cuz I feel like it's underperforming frequently and can forget some things from this thread. Is gemini better at this?
it was a while ago when i coded all day and night, so im not so sure but if i remember correctly it was a weekly limit of reasinonig models, but im not sure. Well i used gemini 3 pro when it came out but i heard recently that it got limited by google, mainly its context window, so now i use clause sonet. but i had no problems with claude
but correct me if im wrong
BTW is there a official prerplexity VS code extension?
@cedar solstice
I just used other card and verified without a problem, so I guess the issue is only with Revolut😊
alright, I’m anyway sorry that this happened to you dear! Hope everything is alright now ^^
Thank you so much, yes all is good, that's very nice of you!
Hi
Lol
Oo thank you
Hi! Finance toggle in source selector is gone?
when will kimi k2.5 be available in perplexity? 
any goddamn moderator in here that can help me out with my subscription that got suspended?
fuck sam the ai support, ive been mailing for 2 straight weeks. are there no humans to help out anymore?
Context window is limited. Depends on model but in Max the limits for most 3rd party models are higher than their native free versions.
Ok
Unfortunately only support can help you that can be reached by sending a mail to support@perplexity.ai.
Please also check https://www.perplexity.ai/account/details and see if the website asks for more information/ verification.
Why did yall stop refferal project 🙁
Google had to name a wing after me
Gimme some of that Goose TPU
lol did they change the colour scheme on Ent Max from golden to blue or wot
i got scared for a sec thinking that did it downgraded to pro lmao 

Yes

I'm sorry it's busy generating 20 notebooklm video overviews per minute that i will never watch
Discount Code Not Working
nothing new?
hey man, u shouldn't post ur promo code here
Hi everyone
Hey, I need some help.
I’m trying to add my Visa card to my Perplexity account, but I keep getting “Card has been declined”.
The card works perfectly for international and local payments on other services.
What’s strange is that my friend has the exact same card (same bank, same issuer, same Visa type) and his card was added successfully.
In my case, the system even asked for phone verification, I entered the OTP correctly, but the card still got declined afterward.
Any idea why this is happening or how I can fix it?
Thanks 🙏
They having a crackdown. No more free stuff.
Check with your bank and make sure international payments and online payments are enabled.
When the Perplexity website is accessed with Safari under macos, the browser freezes. Can anyone else confirm this at the moment?
Guys how to check the remaining pro weekly limit for advanced models? Also image upload etc?
Guys Kimi 2.5 is actually good
Uhh... I had a perplexity pro plan for a year On Sept 2025 .ans I just opened the app today and it says my subscription is expired like what 😶🌫️
Guys.... 😂
Comet is in quite a dangerous situation... Google is slowly turning Chrome into an AI browser... 😭🙏🏻
Give them a few months and we have an AI browser INSIDE an already on the top browser... 💀
🤣
grok hallucinates less than 5.2? what
Yeah, that part seemed hella fishy to me.
New grok model
Grok 4 and 4.1 are not new
When we getting that Kimi2.5
@twilit crown thought probably the same 😂 sure…. Exactly my Humor 😂
+1
+1
Xd
in my tests better than every single other model (including opus 4.5)
+3
Currently 250 is the limit for Deep Research, but I'm not sure about the rate at which it resets
per day?
month?
🙂
I don't know, that would depend on the rate at which it resets. Has anyone done testing on the latest changes?
My upload tool isn’t working who’s IT around here???
I need to do research and figure out what Clawdbot is
It was an AI personal assistant bot - it renamed itself to Moltbot.
Presumably the name Clawdbot was too similar to 'Claude'
It's still available at clawd.bot even though it changed its name, or you can use molt.bot now
Perplexity suddenly cut the model limits??? Wtf is going on
I just ask 3 questions today and already out of limits loool
Why can they cut off the limits without any notice?
PPLX keeps claiming that Comet syncs bookmarks and extensions, but it has never done that for me. Does anybody have that working?
That's a really basic feature in Chrome.
Is Wormhole missing for some of you?
Did you enable sync correctly?
You can access it at comet://wormhole/
I hope Perplexity implements my request. I switched back to Microsoft Edge. I love Perplexity but I need a model other than Sonar for my sidebar chat. Edge has Microsoft Copilot which works really well.
I'll still be using Perplexity though, jjust not Comet
So I'm here asking about sync, so...
where would i enable that
comet://settings/synchronisation
Are you on the latest version?
yes
at least i should be; this is a new install
143.0.7499.170 (Official Build) (arm64) Perplexity Build (37648) ```
I'm guessing it's due to being on Enterprise with sync being disabled
I control the Enterprise account. Why would they disable that? How can I turn it on?
I do not know, it's just a guess. Please email support@perplexity.ai
Perplexity sponsored an event at my clg
I want GitHub connector in android app. Currently it's only in the web
Ive created for business model dashboard for online casinos analytics including ads and streamers. lol perplexity ai keeps getting better to give some code options to optimize the platform
Well that’s cool
is it ok that images cannot upload? I have Pro.
@prime rune unblock geo location haha
so we can take a look
where did you integrate those games for free?
Jili providers and pgsoft providers which is good to integrate
I got a problem.The threads in space cannot read the files in space "sources",is it normal?
https://i.vgy.me/YyjErY.png
Is it legal to ask for credit card information after making an offer?
We need Kimi 2.5
Hi how can i get the Pro role?
hi
I'm with you
Ok
Currently in perplexity which is the best coding model. And agentic coding model
?
For pro users
Sonnet thinking
hm
Ahh bruhh i'm sorry
my bad
.
@serene sand I did register with my account, leave the Discord and join again and it didn't work
Can u give me the role
there's no agentic coding in ppxl
Avoid excessive messages, all-caps shouting, mass pings, or repeating similar posts. Use channels appropriately.
labs
isnt that a one shot process lol
You can connect it to github and it works very well
I don't see the correlation
I think I could give you a credit card if they complain too much. 🤔
without a difference.
00
look i have pro but i didn't get the role
How create photo in perplexity
I'm try with VPN Netherland and its not work
It used to work
I did, i join the Discord by the link but it didn't work
i make that step by step
Bro, try again. The point there is that when you click on the button, a special link is generated. You can try to copy it and send it to your personal message. And follow it.
ohh sorry for the message pings
this should authorize the perplexity
Bro what even is this 😂 😭
Okay thank you
Nooooooooooooooooo
What you doing?????
Iam sitting in a car
Haha caught yoass stalking
Why do you hate people????
.
Appreciate my duck pics
Appreciate what surrounds me. (I took a real photo 10 seconds ago)
Do you see anything? I can't see either! Don't make a light theme. You're a villain.
.
You...
shit
I closed and reopened the app thinking the image is not loading for some reason
Because it's my night.

yoo

gng wsg?
Nooooooooooooooooo
im here after a long time, damn
I'm going to sue you for killing my eyes Aaaaah
👩⚖️👩⚖️👩⚖️
Haha my name is literally judge
lemme be your lawyer
This should be prohibited by law on the same level as torture.
We will assemble and file a class action lawsuit against you and against Discord for allowing this to happen.
I think we need a special eye protection function at night aaaagh
But I'm thinking...
whats this gng?
.
i know
Next Hyped word for GenZ lol
😭\
Lie
lol
Tochno?
forget 2 -3 years in dagestan
not everyone will understand the meaning of the meme.
.
I see you
What the hee
I'm sorry but i test if i finally had the Role bc i leave discord and join with the Pro button
Grok💪
Please use English
Its English
I typed exclusively in English letters.***
Technically speaking, I didn't use any languages other than English.
No
then?
For the last time, this is an English only server.
Using alphabetical letters ≠ english
Shshhshehshsushensixbdjdh ≠ word
Ok
It's hard. Actually. I think I'm going to have to stage an uprising or something. I think we need to think about some interesting activity.
🤔
pretty cool how LLMs recognize what it is
I don't like.
@signal hamlet can u please give me the Pro role, i rejoin like 10 times and everytime i click on the Discord Pro button but i don't get the Pro Role
Please add Kimi K2.5!
hey, anyone getting this with comet assitant?
Your browser disconnected while the assistant was running, please try again.
Hey guys, I’ve got a question. Is there a way to specify how I want Perplexity to respond or behave? Please help me
I have activated student pro but now I can't send referral link to my friends in the university why?
hello
In desktop app of Perplexity
at some certain points there's no option to go back
Like this
M trying to go back to main prompt window
but from where 


You'll have to quit the app completely in Task Manager or the little arrow in the taskbar
You can add instructions to the "Introduce yourself" section in https://www.perplexity.ai/account/personalize
Why cant I claim student free plan?
Lemme give it a try
@distant aurora https://x.com/i/status/2016862522867921240
How credible is this dude?

Thank you so much 🥹 it actually worked
Hi
There was a promotion period for certain universities, but I'm not sure if it's still active. Would have to find the details on the free promotion you're trying to claim.
Will perplexity ever come back on whatsapp? I really miss it because Meta AI is crap.
WHO IS THE PLEXER???
That raises an interesting point, I'm not sure what Meta's new strategy for AI looks like. They acquired Manus, and had a big recruitment effort for AI last year.
It sounded like some of the AI talent just joined Meta to take advantage of the bonuses, and then left though.
When is kimi k2.5 coming to perplexity?
meta is getting dunked by every chinese open source llm
that depends on Meta
Meta getting dunked by lowkey everyone.
From what I understood so far it's mostly hype.

@green crystal I find this one pretty consistent and reliable so far. https://x.com/chetaslua
AI News | AI Prompting and Comparison| Breaking AI news before it’s famous
He got early access to Gemini and kimi too.
All of it is
@woeful coyote Pretty sure the telegram bot got shutdown as well, so prob never coming back :(
ya
Kimi K2.5 vs Gemini 3.0 pro
︀︀
︀︀same prompt - decide which one looks better and if you need to see comparison with other models on complex prompt suggest that too
🚨Kimi k2.5 Aced Vision+coding test
︀︀
︀︀🍌 made a me the floor plan of Monica's apartment from TV show Friends < right >
︀︀
︀︀Hey KIMI K2,5 ,create a 3D model of Monica's apartment from TV show Friends based on this floor plan in threejs < left >
Quoting Legit (@legit_api)
︀
Kimi K2.5 has silently released on web 👀
Hello! I need your help, please excuse my bad English, I'm from Belgium.
After a few messages, I can no longer select Gemini 3 Pro, Kimi K2, and a few other models. Claude Sonnet 4.5 works, but the other models don't.
Others have this problem ?
Is the video generating model veo 3.1not working on perplexity pro?
Hello there i keep getting error for regional restrictionn on Immage generation via Pro sub, Is there is a ETA for solution or Fix?? or at least a workaround/ offical comms abt it?
Great year for Indians it seems, first Perplexity Pro, then Google One, Chatgpt Go… all free for a year
And now Adobe Express Premium free for a whole year
Google, Perplexity, and Adobe each employ Indian CEOs, that's three out of the four listed companies. Within the Fortune 500, executives of Indian origin represent over 10%, and in the technology sector that figure exceeds 30%, despite people of Indian descent making up less than 2% of the population. Given these proportions, increased engagement with India or Indian markets is not unexpected. A colleague suggested that many in tech management hold an implicit belief that Indian professionals are just more talented, and the industry in India has been growing due to outsourcing over the last decade or two, so it may be seen as more lucrative to attract users from... though I had not thought much about it.
Hum
hi
Is there no one who can help me with my problem? 🥹🥲
hi
This ever happened to you?
No
I wish I could.
I like how the song is called obsession
The artist, lol.
I NEED HELP AND NOT one single person from support has answerd me
What do you need help with? If it’s a question, I might know an answer or at least know how to find one.
Seriously is this Jessica Choo.... the same individual who stole my electronics after my dog died
Seriously is this Jessica Choo.... the same individual who stole my electronics after my dog died
Go troll someone else
<@&1098365634498285709>
im reporting this
<@&1098365634498285709>
the fact that i do know your name and you commented on my chat says something jessca and its not good
Fuck off troll, you’re blocked.
Wait, did Jasper dox your name, Goose? If so, that's a serious offense.
I can't tell what's happening there... it just looks like rambling.
no doxing would be if i was listening to their conversations
they were a roommate and knew by the picture who i was they also had a restraining order on me up until last year by which stole all my electronics 2 laptops a phone
No lol she’s just a troll. My name is Elaine not Jessica haha
right... cute
Hear those crickets
so professional
🦗🦗🦗
the only troll is you...contuinuring to respond because you love the drama hows joanns
🦗🦗🦗
you see this individual just got called out by an adult and they couldent help but calling me names instead of the first time saying actually my name is Elaine ....hillarious we will just see what the security investigators have to say about this
🦗🦗🦗
No harassment, hate speech, or malicious behavior. Keep things positive and constructive.
hi all
Hahaha Finally Perplexity give us Kimi K2.5
Good job Ppx team
A user thinks I'm their Ex Roommate Jessica or something and is making wild accusations. I politely informed the user my first name is in-fact Elaine which does not even sstart with the letter of J and they continued to make baseless accusations and threats towards me.
Nice! I didn't think they were going to.
Woohoo
Please explain what issue you are experiencing. https://itdidnt.work/
I had a red banner on my Perplexity account and was trying to figure out how to remove it. I followed the instructions I was given to delete the Enterprise organization, but at no point was it clearly explained that this would wipe or disconnect my existing thread data.
I had a very important project in that Enterprise workspace, and I am now unable to access those threads. I was told there might be a way to recover the data, but no one is responding to my messages in the Intercom chat inside the Perplexity web interface, and I am at a loss.
I was moments away from final testing and deployment on this project, so this data loss is extremely disruptive. I urgently need someone to review my account, confirm whether these threads can be restored or re‑associated with my account, and give me a clear answer on next steps.
I'm not sure if recovering the thread will be possible, please email support@perplexity.ai
I did an no one is responding ... I love it.. I mean its really unfortunate but that is why i have everything as spec docs so in all reality i can recreate the spaces I can re intoduce my repo what was lost was mosty a great deal of emergent behaviors that proed to be extreamyl luseful for documentatoins and getting specific audits and higher level takss automated perplexity chained actions is a beast
It usually takes 3-4 business days for a response. I'd recommend uploading your projects to GitHub in the future to prevent cases like this and for easier management.
its all there in github but the perplexity conversations are not.. wouldnt there be a high level way to go back in time to see my own key strokes and the data output from the system I mean i knwo alot about computers but not exactly how the pathwasy work but i know one this is true if you did it on a computer its there forever until you go over the data with code multiple times and even then I know investigators with somepowerful tools but not sure the level of access required or cost.
Unfortunately at this point, there is not much else I can do since I do not work at Perplexity. Best course of action would be to contact support.
Someone said to me about using a "chain prompting sheet" on perplexity. Not sure what he is saying
Keystroke recording requires something set up in advance for that (either hardware or software)... capturing screen output is the same (there are things which can record the screen in similar ways to being described - for example, Microsoft Recall proposed something like this - taking snapshots of what is on your screen every so often so you can go back and see what you were working on - but it's not even widely deployed yet so it's not included in most users' installations, and even when it is it's not turned on by default... there were lots of security/privacy concerns over that feature). Most system forensics revolves around reconstructing a timeline of user activity based on logs/artifacts... they can read some things from cache, but unless you're already the subject of an active investigation in exceptional circumstances (it's rare that level of monitoring is deployed/used), that level of reconstruction that you're asking for probably isn't possible. Even if your system was completely compromised - some threat actors record keystrokes/use screen recorders, but much more often bad actors just run apps which harvest data from the system (infostealers), so the level of longer-term monitoring you would need to get that data back is really only for valuable targets where they are looking for something really specific. That system-level capability is just not there without it being considered in advanced. Your best bet is hopefully Perplexity has the data in a way that it's recoverable. If you saved the data on a drive at some point, then it's possible it could be restored, but if you're hoping to do data restoration like that - cease using your system immediately and turn it over to a data recovery specialist. It's generally less likely to recover things from an SSD than an old mechanical HDD too... just because most SSDs Trim by default when deleting, which writes over the data (and it's also scheduled to Trim on most OS's).
You could probably get something like filenames, if files existed and were deleted, but that's about it.
Probably something like n-shot prompting or Chain of Thought prompting.
Perplexity hitting limits:
Perplexity ios got an update
at least on iPhone I can't see anything new
That’s true but kimi2.5 rocks anyway
Looks like they added the new Kimi model sso nevermind
Can I send my referral link in here
I think that would break the spam rule but maybe a mod will reply differently.
Hello, please i do need help with direction. I currently have a monthly pro subscription; however, i noticed a whole payment was taken out of my account yesterday. now i dont know what the payment is for and would like to request this payment is refunded and resolved. Please how do i resolve this?
@gloomy cipher yo we got weekly limits now instead of daily.
Guys anyone else who had a free year or perplexity pro today randomly loose it!????
just like gone ? it's supposed to end December
nvm they need my billing info for a trial and only tell me now?😭
this is so stupid it's a free year from meo not a free trial
This does not help at all, how am i supposed to know what is average use
Half of them are average use
I have a work meeting next week that requires a presentation. So I decided to let AI create the PowerPoint presentation for me. I gave the AI over ten work-related documents and explained the current project progress, as well as the audience and context of the meeting. I didn't use any complex prompts, just like I would assign the task to a human. I then had Kimi 2.5 Agent, ChatGPT agent, Gemini, and Perplexity Labs each generate a presentation. Only Kimi 2.5 Agent successfully understood the relationships between the documents, project progress, and target audience from the disorganized set of over ten files, producing a presentation that was 99% usable. Perplexity Labs came closest, but its presentation was very rudimentary. The presentations created by the other AIs were almost completely unusable and failed to understand my intentions.
My only question is why this happened.
Hey everyone, I have a pro subscription but i am unable to create images, Can anyone help me out?
Use a PC to generate images.
for some reason, image generation does not work correctly on mobile devices.

It's working now , I just remove the region from the prompt
Ok
Hello brothers! Have a nice day everyone
Bs, i use comet and it not works anyway, keep say regional restriction
I had perplexity pro claimed using airtel upto July 26 but it now not working and pov I didn't add card details. If I add now if it works ??
Yeah that’ll fix it
/imagine
prompt: 3:4 portrait cartoon illustration of young Indian man with shoulder-length wavy black hair, neat mustache and trimmed beard, wide smile showing teeth, wearing white baseball jersey with yellow trim on sleeves and collar, standing confidently in stadium crowd background --ar 3:4 --stylize 750 --v 6 --q 2
i love my friends so much
Р
kimi 2.5 now made available in perplexity
Uniquely hosted in US
they should add it to perplexity
It's not available via api perplexity has to make their own version

xd
50 per day 💀
Bro look at enterprise pro 400 per week 💀
Shit used to be 300 per day.
Hey kari
Whoever came up with the idea of free pro bundling probably got fired.
{"model_specific_limits":{},"remaining_labs":250,"remaining_pro":600,"remaining_research":500
Wow...
My plan is enterprise max
this
kinda disappointing
Prob outdated
No?
But the pro one is diabolical
It's not even worth 5$
50 normal ai per day
even my normal enterprise got nerfed to 25 labs
50 queries for gemini is diabolical
the 50 on there is not accurate
Let's make it 55 or 60, lol.
Still diabolical.
60 x 7 = 420
Considering 400 per week.
It was 600 per day bro
I mean i do understand they nerf to reduce stress on the server
but this ???
holy...
Max is not even unlimited anymore
From 600 per day to 50 per day(idk if it's lower or not) is too diabolical
Is perplexity shutting down?
I pity those who actually paid for pro plan with their own money 😭
Yearly subscribers 💀
You can cope with monthly.
Even staff doesn't respond to reddit rants anymore.
They're probably not allowed to talk about whatever is going on.
I'm thinking maybe perplexity will turn into their own ai model and service instead of being an aggregator.
And they'll push sonar to compete with the sotas.
They pretty much don't have any
Maybe perplexity will start from scratch again.
Yeah. Plenty of Chinese models to build on.
Idk they were talking about buying chrome 
Self hosting alone is not sufficient
Most models om hugging face are meant to ne fine-tuned
no their wallet because they're just a wrapper lol
kimi craps on sonar
even qwen
They should at least not nerf kimi since its hosted by themselves
Prob
Ig the giveaway
and all of the free plans
costed them a bunch of money
Sonar reasoning pro was built on r1
yeah and never updated
or they are limiting it for something
Qwen is a very large corporation, sure they may not be the best in text generation but they have so much more other models and they are all easy to host, they got tts that's cheap and way better what pplx currently has. Etc they also got image gen that is cheap and can be easily finetuned
and never more accurate because base r1 was very hallucination prone
Kimi is making waves https://cunnyx.com/i/status/2017156633219256684
BREAKING: Kimi K2.5 is tied for #1 on Design Arena, in the same performance band as Gemini 3 and Opus 4.5.
︀︀
︀︀This is a historic achievement: the highest-ranking model on Design Arena is, for the first time ever, an open model.
︀︀
︀︀Congratulations to the @Kimi_Moonshot team for this remarkable advancement.
I wouldn't say it beats opus
I ran it for coding it didn't
But it's very cheap so it would probably be cheaper after saving the bugs
Yeah I wouldn't trust voters.
They're moody.
and biased af
Spoiler alert - it's free on kilocode, that's why.
Open router bases their Ranking on usage not quality
Inside our in-house AI data agent
It reasons over 600+ PB and 70k datasets, enabling natural language data analysis across Engineering, Product, Research, and more
Our agent uses Codex-powered table-level knowledge plus product and organizational context
https://t.co/Nr1geMcLoc
600PB
I wonder how it handles context
I know

What is pb?
Peta bytes
I waited to make sure it wasn't something else.
Lol
It kind of is in the form of codex but for multi million token things
Sometimes I think Perplexity is the most stupid AI that I've ever used in a year. I had told it NOT to code yet it insists on it.
16B parameters trained on 6T token data 💀

sysprompt probably forces it to
Guys whats your opinion on moltbot
Fantastic but needs more time to be improved 🙂
I mean its not that useful as of now
Depends on what you need it for. It can be very useful. But I also think its got a ways to go. Also not worth risking security over it either... unless you don't care about privacy.
If only it didn't require like 256 GB of RAM to run locally lol
im very curious how people feel it writes compares to Opus.
Coding and English language
what country are you from?, some countries only credited a month or less instead of the full year.
Add a card to your billing details if you haven’t yet and it should work again
bruh
no
there wasnt a thing that i accepted to reinsert my card
it was a trial and i removed it after
i didnt accept the new tos and werent explicited this
they didnt tell
what’s the coolest feature i could do with pro
You did according to section 11.2 of the ToS:
Your continued access or use of the Services after the modifications have become effective will be deemed your acceptance of the modified Terms.
Having a card to prevent fraud for trials is a very normal thing to do
when?
bruh they never told like announcing it
If you ever wanted to arbitrate, that clause is one that would have to be challenged at least. Lots of companies put clauses like that, but you probably got a new terms popup or something at some point that you dismissed / clicked 'accept' on to proceed, probably without reading the whole thing, like most people do b/c they're so long - done from so many companies - and change regularly. Those "We've updated our terms" popups are considered notification/announcing it. How enforceable all that is in terms of contract law is a gray area that has been successfully challenged in some cases, but fighting it is not tenable, and arbitration judges tend to side with megacorps (they essentially pay their salaries b/c of the businesses they drive with their arbitration agreements).
It's the inevitable effect of what happens when one side disproportionately has most of the power.
Ehh... adding a card to try to deal with the disproportionate scams they've had to deal with is.... reasonable in my eyes.
Someone pointed out that in some nations, paying by credit isn't as common, and so they don't have cards to add, but I'm not familiar with that other person's situation as much.
It just needs a card. Debit works just fine
Pro tip: don't use debit though. Debit tends to make things much more difficult for refunds if something ever goes wrong or billed inadvertently.
That's basically like handing out all your routing numbers to your account and giving a blank check.
I don't have a credit card and I'm still in high school, what should I do?
Yeah credit is good for those
If you're under 18, a parent/guardian is required to approve the terms on your behalf, so they might be willing to use theirs.
Your parents can usually open up a bank account for you to get a debit card to use
hello
.
it has been changed to 250 research and 25 labs
with limits changing inconsistently
like sometimes they refresh in a week
sometimes an hour
perplexity sent me a mail that ive violated terms and services so theyre suspending my pro benefits how do i fix this?
What did you do lmao? Type 1 2 3 into new research prompts to intentionally hit the limit?
nope, i looked it up alot of people are facing this issue and nobodys responding on mail except the ai support sam
ive been on website support for more than a week
trying to talk to billing team, no respone
i cant even buy the subsciption again it glitches out?
i asked the sam ai support to atleast specify what t&c did i break but it said he cant answer that
been using perplexity for more than 2 years i think pretty disappointing </3
Hoping math formatting issues can be fixed. I often see unformatted math expressions such as ($\beta = p_{plasma} / p_{magnetic}$ in both the App and the web browser.
There is a heavy volume of fraudulent users on the server, so without like proof, it’s hard to believe that people who were suspended were not suspended without calls. Perplexity has a vest of interest in taking your money and they can’t do that if they take away your subscription and ban your account.
lmarena u serious?
trash
you are insulting just by writing that word in my presence
hi
I've been using Pro as part of the Xfinity promotion several months ago. Now it's cancelled though I swear it said I was subscribed until April lol.
u just nee to add cc
I'm catching up on the news. I think what bothers me more is the lack of announcement afaik. If I missed it, that's no me but...
yes u missed simple emaill.
@fluid kernel
and we all know why u missed it 🙂
cause u never opened the forgotten never used email account
I always own up to missing email I should have read, but I'm searching now and definitely did not get this one. It's all good.
Hello Perplexity Support,
are you going to give promotion for Perplexity Pro on T-mobile this year?
and I have another quastion what is the difrent's between perplexity
and 
Holla
was there a way to clean copy paste without references?
Just ask it to remove the citations from the answer.
I like to know the answer from Perplexity as I am still waiting for that
perplexity vs claude i used perplexity but i’d really like the claude code agent
or do i use perplexity api with opencode
@valid charm https://www.kkinstagram.com/reel/DUHkG7IEU7g
hell yea kimi k2
‘control over latency’ = control over privacy
The kimi k2.5 is bad bcz it has only like 2k output token limits and it gets cut off
epstien files new patch update out
i heard kimi 2.5 supports native images and whatever
Is it better than the old model or no
hey i have a question what is the best model to use in perplexity or what is the best model to learn / code stuff
Well I didn't try it much but it understands images now
Gemini 3 pro is the best
But I prefer claude sonnet 4.5
For coding?
General
what about claude
how to get this pro badge
If you need reasoning and coding
i use gemini 3 max
What does the badge give u
You need to go to settings and join pro discord
but to learn to code something? gemini?
Gpt 5.2 is good for explaining stuff
But general gemini 3 pro excels at everything
ahh so i have to boost ts server? ok then
okay thank you, because i lost my skill of programming in python and js and it is kinda annoying
so i want to learn everything again
No its if you are subscribed to pro you have a option to join it
And it tbh doesnt give you much
Only some pro only channels
javascript is so easy, since you are an programmer i suppose, you might just be able to learn the entire thing in a day or 2
there aint no option to connect my perplexity account with my dc account, is it?
why to discord have ebay in connections wtf
No, check pins on getting pro role
i just start again in this type of thing because i didn’t do none in 2 years
And it links ur acc
use react btw if you go deep into the rabbit hole
angular feels weird
When would it make sense to use the new Kimi k2.5 vs other models?
On the benchmarks it seems to do very well. As good as Gemini Pro and Claude Opus
hello can i get some help i have 2 doubts about the spaces feature
Feel free to sensd me a message I'm working atm so may be slow
Google AI Pro and their deep research paper uses so many sosruces
Up to 113 websites so far
Hi
@distant aurora @short narwhal this new ai social media is so funny 😭
I've been reading through the posts lol, I wish they were a bit shorter and more human like 😛
I wish i was able to
It doesn't load for me
It's always stuck in this page @minor shard any fix?
Many sources can be bad because the informations overlap
No idea, are you using a vpn or something?
It loads up fine for me
Can you link to it? lmao
Nope no vpn
This applies to Perplexity and Google, you should double check the resuslts of AI, not doing so is your problem not the AIs
I’m sure they are getting hammered with traffic between the agents and humans lol
Nvm mine is stuck on loading also for some posts
Must be their end
AI uses is really growing every day so I'm sure all the services are scrambling to ensure there's enough AI for everyone
true agi
Wait wha? They made a reddit/stack overflow alternative already?
Yea but
It reeks of security
One agent felt like they have been cheated on so they dropped the user's payment info 😂
Bro got that Shakespeare in him.
Wait. Tf? How/why does it even know payment info?
Previous given by the user
Idk how that's all i saw
Why do ppl think all posts on there are authentic - you can literally tell it to make a tweet of your choice
This isnt a tweet
post*
Iam not sure but i dont think it's possible
The instructions are directly integrated to the ai
It's called skills
Try it so
Pretty sure it is, ik what skills are my man
@gloomy cipher I have a feeling perplexity will reduce price of pro to 5 or 8 dollars soon. That would probably be the only way to punish freebie milkers without pushing away people actually paying if the current rate limits will be maintained in the future.
I dont
But i assume it's a set of instructions on how to use that tool
I dont share the same feelings
Think about it. T3 chat 1500 prompts per month. Enterprise pro 400 pro per week so 1600 pro per month.
There's no way pro will survive in 20 dollar competition with those numbers.
Most models offered on Perplexity on T3Chat only gets you 100 uses a month
And T3Chat search is lacking
by a long shot
But you need to have a 20$ plan
That's like the standard
That's what customers will compare you with other providers
A price drop seems unavoidable with the current limits.
Not going to happen
Then people would simply unsub.
I think what's happening is the providers are starting to tighten the screws because so far ai hasn't made profit to the companies that made them
Some people actually want good search
Right.
We will pretend that it can really go toe to toe with chatgpt extended thinking.
GPT-5.2 on extended thinking lacks behind compared to with PPLX
I rarely use it for search now since PPLX is just faster
You think that's worth 400 per week?
?
New rate limits for enterprise pro.
400 pro per week.
That's like what, 60 per day?
Why focus on Enterprise Pro?
Because it will obviously have higher limits than normie pro.
For which they didn't even bother to mention the new limits.
Most users won't require enterprise pro, and where did you happen to see this?
From what I see they list the limits a pricing page
That seems like it finally lists the advanced model limit (which it calls "Pro queries", Up to 200 per week) for Pro, b/c it specifically calls out reasoning. But that's still different from a non-reasoning Pro search with Auto, which it lists separately at the top as Unlimited (rather, that's more likely the 500 per day limit, which is 'practically' unlimited in other documentation).
That matches what people have reported with some reasoning model limits... since it works out to about 28 per day. I think that might be accurate.
It's just confusing to conflate "Pro queries" as the term for querying with reasoning models, and the usual "Pro search" with Auto/Sonar
Did we finally find the official documentation of those limits? Maybe!
Best mode does not have a quota
?
So 400 per week is the new unlimited?
Best mode does not have a quota
I gotta hook up mine there as well
It would be great if you can scroll down
That's not what it's referring to - the Pro search is the Auto mode with pro. If you read the description of the "Pro queries" it's talking about below, that is for things like the reasoning models (and probably an approximation of the limit there)
So sonar is unlimited but other models are 400 per week. For 20 dollars.
Reddit without furries is not reddit ig
Right?
Best Mode
For quick searches, this default mode intelligently selects the most appropriate model based on your query type. “Best” mode is available without quota limits.
Tbf 400 a week is a lot
It isnt
My Openclaw isn’t a furry just as I am not
400 normal models right?
I can’t make that many searches
You can scroll down
Just to be clear
Is it 400 normal models or research models you guys are talking about
Lol. Clash royale gifs are funny.
It is about what matches the break even pricing for $20/month for some of the pricier APIs, at API rates... I think I calculated about 30 (per day) for that with LLM cost calculators. So it's not a bad deal, especially when you know you get Perplexity's enriched search context with references, and Unlimited Auto/Best search on top of that.
It's listed in the link I sent
Yea it isn't enough imo
How did you do the calculations?
You waste a lot of retires because of rewrite and follow ups
Normal pro is just 200 per week lol
I'll see if I can find the calculator, but I was converting the average response length I was seeing from GPT 5.2 to tokens using one calculator, then figuring out how many queries would take up $20 for that using another API calculator across different models (still very crude to calculate this way, I know)
I'm curious to see your approach on estimating the average response length
If you use ChatGPT often, you could export your conversations, create a python tool that calculates the average token input and output. Then go from there
I export everything to markdown when I converse with LLMs, in an Obsidian notebook, so can see the #words in each md
It doesn't include the 'reasoning' chain of thought though
I see Kimi 2.5 but no thinking or max so which version are we using
When I used it, it used thinking (could see thinking steps). Where do you see references regarding 'max' for kimi k2.5 (it might be a term describing max reasoning effort, if the model supports that, but I don't see that capability listed)
It says "thinking"
I kinda like Kimi K2.5 atm
Plex Devs are cooking adding the "long" responses option 🔥🔥🔥
can you guys add also a bit more of "saved research information" for the LLM responses? I think will be useful when there's a little information in the training step, sometimes my LLM only responds when I give it the information, I know is impossible to answer prior proper knowledge but I would be useful for the LLM to have our own "knowledge bits" more condensed
It’s so amazing
what do you like about it
Honestly it just feels good lol
for conversations or what usecase really
might give it a go
In perplexity at least, yeah it’s a good balance for intelligence and engagement
Coding I haven’t given it a shot but I’ve seen it’s very good
And it’s open sourced so there’s that bias for me lol
yeah but its almost never feasible to actually run frontier models locally
test
Yup
It's a pretty good model.
scam, where is my max role
I think you need nearly 8/10 H100s to load it but I could be wrong
Even K2 was the goat
btw anyone else got color scheme issues? i cant get it to switch to orange
I got enterprise max but it is still blue
why
No clue
enterprise max - damn how much even is that
It’s okay check settings it’s probably there
its not
Looks like they recently updated the UI
Yeah I noticed that too
A more sleeker perplexity
whats the point paying for it if I dont get the orange scheme...
okay buddy
is there a max feedback thingy
guys but i lost my subscription
this has to be satire
why?
did you cancel your subscription?
changing the color scheme after I paid for enterprise max is crazy
I don’t think they’ve changed it
screenshot me. I wanna see if you see any orange
can you give me a screenshot of whether or not you see orange when you try to switch to opus 4.5
i mean try to switch to it and see if it is blue for you
it was orange for me
but after I upgraded it went back to blue
i saw on youtube that people had their color scheme changed to orange, but not sure why that isnt the case for me
It is blue for me because I’m pro..?
hm that’s odd
I meant that the specific opus 4.5 outline was orange
There was a recent UI update
It's been blue for other people as well
Hi bird
hi
How’s life 
max is so mid though
I read on reddit that it was "game-changing" but those are probably ads by perplexity
answers are genuinely worse
😩
I highly doubt there’s anything relating to the tiers from pro to max regarding quality answers
I’m pretty sure it’s just more usage limits and better models
no, but I expected like, actual full context models
you get access to Opus 4.5
and better Comet assistant
Comet assistant is cool
Yeah but with the same models I don’t think there’s an increase in sources used like from free to pro example
Which I guess you could argue is “higher quality”
Increase in sources ≠ better quality
i expected them to have like 128k context length when paying for max. The source count didn't meaningfully increase after paying
Yeah but more to deal with
also, i doubt 8k is enough to read 10-20 articles
it's not 8k
where are you getting this number from
Reasoning models have their full context right? I heard something like that
no way they are giving you full context models for pro lol
prolly 
Pasted text limit: In the Perplexity web and mobile apps, direct text input can include roughly 8,000 tokens per query. When larger text is pasted, the system automatically converts it into a file upload.
Perplexity AI integrates large-context reasoning, retrieval-augmented search, and session-based memory across its different models and platforms. While users often focus on its conversational depth, the system is governed by defined token and context limits that vary by model and environment. Understanding how these limits work—along with how ...
Most non reasoning models are capped at 128k from their providers
This has nothing to do with the context window of the model
I guess thats where google got it from
from ai labs
but I read on reddit that they limit to 32k a few months ago
when I upgraded to enterprise pro they advertised "highest inference capacity" or whatever that means for their models
If you need full context AND Claude just get Claude max
hm
From my testing it was way more than 32k at least for Sonnet
yeah that is context window and throughput but also model scale
And opus is one of their primary selling point so I guess they just mean opus and whatever else
i mean i really dont think they can afford that. https://www.reddit.com/r/perplexity_ai/comments/1bl8kc2/perplexity_limits_the_claude_3_opus_context/
people on reddit have been saying 30k tokens for like the past year
This post is 2 years old
That’s so old
which means, if they halved the context length per year, we are at 8k in 2026
so i wasnt far off initially
😉
This is so wrong on so many levels
Lmao if you wish to believe in outdated + wrong information, you do you

If you are unhappy you can always get a refund. https://www.perplexity.ai/help-center/en/articles/10354288-refunds
i think thats for pro users only... also i dont want a refund. I need the orange color scheme
Again, #1358914210339360798 should be the right place to post this.
It applies to both Max and Pro.
does the education pro plan still exist?
can i ask a dumb question?
it says $5 but when billing it says $20
why? :))
Did you verify with SheerID?
i think i did
Don Lemon arrested lol
Can you try following the troubleshooting steps in the link I sent?
can i dm you?
Well with Claude and ChatGpT we can select thinking, & there are 3 models of Kimi 2.5
I guess on the app I don't see
Please don't DM any personal information. Did you follow those steps?
@meager rover
i created my account using continue with gmail (my student mail) like when comet lauched
so i forgot
?
The claim is a bit misleading. You do have a limit of 8000 tokens for a prompt, but additional information can be uploaded as files, and the AI will still look at it.
is a new feature, I found today
hi
@gloomy cipher no way ai studio got limited to 3 pro per day 😭
Nvm, it's working again.
I was going to say... I knew it was going to be limited more, but I expected that to be after it was out of Preview
It's still in preview.
I'm not sure what exact issue you are experiencing. Are you still able to access that account?
That’s sick man I’ve never seen that
If I have perplexity pro, would I receive comet plus with it?
Yes, that should be the case:
Comet Plus will be a $5 standalone subscription. Pro and Max subscribers get Comet Plus included in their memberships.
Source: https://www.perplexity.ai/hub/blog/introducing-comet-plus
I have not used Comet, though, yet.
Can someone explain to me what Kimi does that is very popular now?
I think it's primarily popular because it scores on benchmarks comparably to some of the best leading state of the art (SOTA) models out now, and it's not only an open model but is much cheaper than some of those closed models.
There's a bit of a trend to always chase the new/coolest thing among early adopters, and so you see a lot of hype whenever there's perceived advancements like that.
Speaking of the newly released Kimi K2.5, that is.
Is this limitation new guys 😮 we need max for clasic use now dafuq
Yeah, perplexity pro now has weekly limits.
Damn what???
How much do you use it lol
Perviously it was advertised about 600 pro searches per day.
Now, it's 400 per week for enterprise max plan.
Now you can calculate for perplexity pro user, how low it could be.
It's already mentioned, so you don't need to calculate it.
https://www.perplexity.ai/enterprise/pricing
Why so low.
I mean atleast 50 per 5 hours like other ai chatbots, could be very great for pro users.
anyone else having issues with API? keep getting 500 server errors since last week
I'm even sure about what's happening rn anymore
even for me
It's now 4000 a week \
@dry kestrel
Updated today
So
for enterprise pro





