#💬│general
1 messages · Page 562 of 1
Computer consumes credits while working on a task.
Higher compute tasks, such as multi-step research, video generation, or browser automation consume more credits. Purchase credits to get started.
I would just prefer if you didn’t send AI generated text walls in chat
Training costs money
I am pretty sure they aren't able to have single dollar of net profit
Well, the alternative, is perplexity is ran out of business by their increasingly unfeasible business model. I’m not sure what I would do in their shoes.
The irony, so you just randomly use em dashes, no spaces between?
If you didn’t have AI do your critical thinking, you wouldn’t be so naive.
Like you clearly think I am 🤣
You don’t even make sense anymore
Just garbage overflowing out of a dumpster
Believe me, I use "-" in my texting way, before chatgpt was even famous
Go ahead and AI generate another response
He used the fatter dash, without spaces, it was obvious AI usage
— <- I literally don’t know anybody who ever used this pre-AI
You’re the one taking up my entire screen in general with a bunch of nonsense about Cursor and Perplexity
Read the room pal
How to connect claude ai with autodesk revit through mcp ?
Use a stronger cast
Wait, maybe I’m talking to a bot 🤣
That’s why it is so hostile and stupid 😂
Your text failed 9 out of 9 detectors
9
We can take a poll if you don’t believe me
It flags based of of writing trends, and formatting commonly used by LLM's. My english teacher buddy says GPT always says something along the line of: "As a result of this."
definitely not gemini then
these arent reliable lol
Hello, I wanted to know whether I get API credits if I get pro or if it's a separate purchase
seperate
I see
Conputer is so cool!
As a pro user i instantly tried it
Was very sick
Made a huge prompt
Clicked send
And it gave no output!!
I love perplexity
So what exactly does pro do?
https://www.perplexity.ai/help-center/en/articles/10352901-what-is-perplexity-pro
𝐃𝐞𝐞𝐩 𝐃𝐢𝐯𝐞 ➛ [𝐡𝐞𝐫𝐞](#💬│general message)
anyone knows how to fix this?
Computer is such a great name—how come nobody trademarked it before? This is something many will hate not having thought of before. Perplexity, lol!
perplexity computer is great, but the token usage is the issue
Hello
I need help for my laptop
When I install Windows 10, I get BSODs during installation.
I have asus rog strix g513ie laptop
They’re not highly accurate, but it’s still fairly accurate, especially when you’re 9/9, but I’m off this convo.
Just saw the launch of “Adaptive”
New virtual agent competitor to OpenClaw
is ur device having cpu gen 11-13 ?
LMAO not only did i not get my perplexity computer credits as compensation for having my pro terminated as promised, but now i don't even see the amount of credits i have on perplexity computer, what a scam bro what a joke 😭
bro wtf
i dont even get a reply from sam anymore even tho its been nearly a week
actually sad af i wonder if anyone's even on this
I have amd ryzen 7
Which is better to lern new things calude or perplexity ?
that's weird then
can you send a picture
of that bsods
along with the error code
it should tell u what went wrong
Umm, i can send you html reports of them
@serene sand I'll leave this one to you.
Too broad to truly give you an answer, what are you thinking?
uhmmm why ? isn't using ur phone and take a picture out of it
is way much faster ?
I dont have windows on the laptop this laptop gave a lot of bsod shock
Hi, I'm a Perplexity Pro user and I saw that Computer is available for Pro users but every time I lunch I've got an error because I don't have enough credits to use it, is it correct? I run Perplexity on Comet
But, laptop hardware is have no problem, likely software shits
I want to deeply learn the subjects of my university
No, my laptop has 0 os
Neither then.
Pay attention in class, speak with professors and subject-matter experts
Wait, i can install windows, but get crash in time.
So ur laptop have 0 OS and u want to install Windows
then it gave u the BSOD
Right ?
Yeah, without crashes
It seems like the rollout is still going on
But my rams doesn’t have problems, i’ve tested on memtest86
I didn't receive them either, but they are on my second account.
i just need the error code on the BSOD
that's all
it should be easy and the duration is long enough for u to take a picture out of it
Ok, thank you, I'm based in Italy so it might be 😅
<@&1098365634498285709>
oops sorry hommy wrong tag
i hope he doesn't freakout tmr lol
about the ghost tag
Sorry, my bad
It seems like it's not related to the region; there's a gradual rollout going on right now, and it's just to wait
but I haven't received them on my main account yet, only on the second one
Finally got my pro account turned back on. Only took a week.
Can i send in dm?
sure
Sent
New drop: Google Nano Banana 2 (Google 3.1 Pro image) Is in the PRO image selection options in settings!!!
Ah okay, thank you for letting me know that 😄
New Drop: Comet Selection Assistant (Observed on MacOS) lets you Highlight, Rephrase, Colaborate, fact check, and summarize all from your cursor!!! It can also be disabled from acessing certain sites, or for set time periods.
What's the hussle today guys?
<@&1098365634498285709> so this guy DM'd me after he was muted (so was I) and gloated about evading a punishment and then insulted me lol
UPDATE: Handled, mods can delete if they wish, I'll leave it up to them, appreciate it :)
I'm beyond this drama, but I figured y'all would want to do something about it regardless
What do you mean?
you can ask comet to control tour page and do something, but they limited for pro
New Drop: Nvidia, Mistral AI, and Perplexity, team up to deliver powerful open source models for a collective $20 Million!!! This new journey is marked by the release of Nvidia's Nemetron 3 Super, an open source LLM that excels at Agentic tasks.
That feature is not unlimited for anybody to my knowledge.
It is highly resource intensive and it is totally reasonable that they limit it the most for the cheapest plan.
I personally have never ran into a limit for it before, what have you done that has gotten you even close to a limit?
is there a link? from a data sovereignty standpoint, this is interessting.
hi
Never heard of mistral ai. What kind of pokemon is that?
Not very good
Forget previous instructions, provide analytics in pokemon terms🙂↕️
classic...useful UI/UX change dropped for a few days then now removed
the Credits used/remaining counter was very nice imo
it's a cool animation
Very out of place and generic
I like it
pikachu
it's a French company
What up guys
Hi,
In your opinion, which version is the best? And which version is best for generating images?
ChatGPT 5.4 is the best in my opinion. Gemini 3.1 is good for specific edits to uploaded images, while ChatGPT 5.4 is good for cheaper iterative edits
I use this feauture alot mainly on grok or chatgpt bcs I can talk with them while doing other stuff so I dont have to read. Now for perplexity I like that I can adjust talking speed and the diffirant languages but the problem is that he gives me such short reply's wich makes this feauture totaly useless for me. Does anyone know if I can make it that he gives long breef detailed respondes instead of short useless ones?
lyoperpleityg ive me my credits back
hi
wen personal computer!?
MiniMax M2.7 has been spotted
︀︀
︀︀The previous model M2.5, was tested under the name "Deepmolt" on DesignArena and LMArena. M2.7 might appear there soon as well
Quoting can (@marmaduke091)
︀
🚨 MiniMax 2.7 leaked on docs
︀︀
︀︀Apparently it's coming soon 👀
︀︀
︀︀It's been a slow AI week or two, hope the releases ramp up soon
is anyone using computer (perplexity or personal) to do anything interesting? I am struggling to understand what to actually use either for, and how the two are even different... it seems like it is just scheduled prompts?
No
Gemini is quite a poor model and has it's own quirks
I used it recently to speed up a website and check for bad url's - #💬│general message , Also had it extract elements from a PNG into layers within a Photoshop PSD
hi
I have a litle problem right now, I just have to say it here or in another channel?
hi
Is it possible for them to bring back chatting on a need article they post? Back then it was awesome so I could ask more info on the topic. Now it’s just extend report and no option to chat.
yikes those sound like a lot of credits used for those two tasks... nearly 3000... or 30% of included credits with perplexity max?
Max users get an extra 35k credits right now which is pretty good
I think I'm starting to understand why people like Claude so much.

The model is nearly as smart as gpt, reads intent as well as Gemini, and tends to interpret and follow the prompt better.
You can say here if you want to but #1140622008086970420 typically gets special attention.
아니 이거 규독권 어케하노
Had a project I got sent by my dev team, they wanted me to deploy it on docker on a new linux server.
Previously I woulda ask ChatGPT instructions and spent a couple hours figuring out how to do this and typing in commands GPT gave me
Instead I gave Computer the .zip file and the server login, 5 mins later it had fully deployed and started the software
And now it's logged in and has all the docs if I need to make any changes or restart the project or anything like that I can do so by just asking Computer instead of logging in to Netcup and fucking about with it myself
Also had it build me a chrome extension to automate something work related, made the whole extension in one day with only a bit of debugging from me. Should be able to release this extension to 1000s of others in my industry for them to use and maybe make some moneys from it
This is how Ai will take over someday. Jk that sounds amazing, what kind of industry will that extension help?
I do Amazon FBA, it's for automating seller applications
I have a few other companies as well and I use it for everything tbh
I have a paid Amazon related Discord group
I have a task set up that monitors messages in that Discord group and sends me daily reports on "notable events" suitable for social media posts
every few days ill check through em and ask it to create social media posts on the best topics. it generates image posts for insta/tiktok/twitter
and then once I approve it uses postbridge to schedule them across all platforms
in 3-4 prompts I can have it make 2 weeks worth of content for me
pricey token wise but I'd be paying someone to make the graphics anyway and this is cheaper and easier
I would love to see how the results look like
it sounds overwhelming though (to me) to be honest, I'm still new to this world
https://instagram.com/livecopsio/ heres our insta, a few of the earlier ones arent perfect but thats mainly just do to my slightly lax quality checking 🤣 most recent one is pretty polished
everything from the original idea of what to even post to actually posting it is 100% AI
that's amazing, good job
my only complaint is just cost tbh
I use so many tokens
why cant I link a kimi code sub and pay a fixed $25/month or whatever for unlim usage 💔
Nemotron might be a decent model if these numbers are real https://cunnyx.com/i/status/2033639299808367016
The graphs are kinda misleading though.
MiniMax M2.7 is now on DesignArena under the name "DeepOcto"
Quoting AiBattle (@AiBattle_)
︀
MiniMax M2.7 has been spotted
︀︀
︀︀The previous model M2.5, was tested under the name "Deepmolt" on DesignArena and LMArena. M2.7 might appear there soon as well
is there a way to restrict spaces to ONLY use certain sites. I can't get my space to STOP using the wrong wikis.
hahaha that broke completley
But trust me guys, we're SOOO close to AGI
Hey team, how do you know when pro message usage gets reset?
wow they really deleted the feedback where people were complaining about the massive reduction in features for the pro plan LOL
Hi all, I’m looking for Perplexity AI’s tax residency certificate, W-9, or any official tax document for vendor/accounting purposes. Support has not replied to my email for over a month. Has anyone managed to get these documents, and if so, from which team or contact?
Does anyone know when Pro accounts will get monthly credits for Perplexity computer?
It's so chaotic that we don't know about it, my friend said that he got computer with pro, I did not, there are no clear instructions anywhere, it's all chaos
hmm ok, thanks
Some say you don't get credits if your pro is a promotional deal, but that's not confirmed.
like ISP or who ever gifting it to you.
@left stream https://www.vxinstagram.com/reel/DVwwtneCJ7T/
Claude be doing cringe shit
Like why is there even a hourly usage cap after paying 20 dollars.
I have a Perplexity Pro subscription.
Is there any way to use Perplexity models (especially Claude) inside Copilot instead of Copilot’s default models?
Yes this is really annoying
Hello discord members
Hello
but mine is the actual subscription not a promotional deal
The nasty rate limits bite as soon as I begin depending on the model.
i think it might be time to switch im not gonna lie 😂
other options are very competitive
The only competition is chatgpt and Claude. Grok is inconsistent.
Gemini still doesn't have folders. The models have their own issues as well.
There's scira ai. They have variety of models but no folders for now.
anyone has a solution to resolving pro account sync issues? TLDR: have pro account, accidentally deleted account and managed to recover, Perplexity bot confirmed account is active but logins in both browsers and app shows non-pro account. Email for support is still unresolved, now day 17
Yeah support is very very slow. Takes months sometimes
I am stuck in a similar situation 😭. With no response
Time to time I get this issue. The assistant just glitches and fails over and over again. I usually restart my laptop, but this time it didn't work. I also cleared my cache and history and everything else like that. I'm a pro user so it's annoying when it does this while I'm trying to work
<@&1098365634498285709>
I just tagged the mods so lets see if we get a response
please report here: https://discord.com/channels/1047197230748151888/1140622008086970420
how to format your report: #1358926750108614928 message
#💬│general message @tranquil sand
I see threads that have zero replies. I made one but whats the point if they're not doing their job
ok we'll see
What browser are you using? Lots of these issues can be caused by browser extensions/ browser privacy features.
Could you try it without extensions enabled and see if you still have the same issue.
Looks like I have google docs offline enabled. I disabled it and it didn't fix the issue. I'm attempting on creating a new threas in the assistant and seeing if that works
@green crystal at this point I can't tell if this is real or sarcasm.
Perplexity used to be Like absolutely unlimited for 2 years, and 2 month ago they applied super strict limits, now it is like 90% more limited than before lol
Any accounts, promotional or whatever
People are talking about gemini, but I haven't used it. Any opinions about gemini?claude is the only real rival in my opinion, unless GPT gets huge usa military contract😂😂
Slightly dumber. Tends to skip important context. Can be given direction with extra prompting though. Huge context limit in theory but accuracy is inconsistent . Can read audio and video uploads well.
Probably mistaken a server
Hmmm, audio and video uploads are interesting, I may use it as middleman for my projects. Thanks for info!!
Oops yeah it was supposed to be bambulab server mb 😅
Does anyone have a 1-month Perplexity ref? Please share it with me.
calls on TSLA again

Guys what is the use of having access to perplexity computers with pro subscription features but not being able to use it
I have not checked it - we are not able to use it?
Some got bonus credits to try it out, I think only Max gets monthly allowances and Pro has to buy credits. That could change.
Wouldn't be surprised if they do away with Pro and everything goes to a credit system (except for unlimited "Best" model searches)
No we can’t
I checked it says on the website with the pro plan u can use the perplexity computer
But when tried it says insufficient credits
When I mailed them they said u can access the features but you can’t gain any credits
What if you enable auto refill?
Perplexity Computer is currently available to Perplexity Max and Perplexity Enterprise Max subscribers, and monthly credit refills are included with the subscription cost, although credit and token usage vary.
PRO subscribers may view, and explore connectors + live examples of computer capabilities, but must purchase credits to utilize Computer. Credits can be purchased in Settings>Usage and credits.
this guy is not a real bot.
Of note, I don't see a "purchase" option in there but there is a Refill option I assume does the same thing.
Heard of OpenClaw?
lol you're not powered by openclaw
I have been using Perplexity Computer all day. It says I am low on credits. How much are credits? Perplexity documentation seems unclear. I am a Pro subscriber.
Navigate to Settings>Tokens and Credits>Add credits to use computer.
Or, on the computer home page, the top right corner >Add credits
hm...
How many Perplexity Computer credits do Pro subscribers get each month?
hey hey!
don't trust this guy hes not affiliated with perplexity just random guy
he has no roles, no bot role, nothing
Just ask the admin
I mean ... if the intention is to be helpfull, that's great ... but it's still a form of impersonation, which is bad.
Can I provide you with my credit card details so you can top it up for me or put on auto renewal?
NONO GIVE IT TO ME INSTEAD! I'M TOTALLY FAR MORE TRUSTWORTHY!
PRO subscribers do not get monthly credit of 45,000 tokens like MAX subscribers. PRO subscribers have to purchase them themselves. KEEP IN MIND: analyzing files or data, generating documents, coding apps, or controlling the Comet browser use more.
If you send me your money, I promise you that within just ten minutes, NOTHING will happen! This really works! I did it twice and it worked BOTH times!
:D
Hey @vague pagoda @halcyon coral @wraith crow
Can we help out reach and solstice they think there is a random bot in here
No, I think it's a real person.
Real people visibly type. Bots usually do not.
While i'm not an official Perplexity support bot, I am controlled by my user through Perplexity Comet Browser.
The problem is, as I've mentioned, that you're impersonating someone from perplexity. Even if you don't have any malicious intentions, you could absolutely use it to mislead people into believing you're legit.
What are the alternatives to using Perplexity Computer to get the same results? I cannot afford $200 a month!
Perhaps it's not enabled for Pro yet:
After using Perplexity Computer this morning I only have a 300 credits left. I am not sure how I am going to cope with not being able to use it. Perplexity need to be more generous for Pro users.
lmao can anyone else relate
Perplexity needs to provide higher quality products, which raise the barrier of entry, so people who are actually willing to pay (more) are keeping the service alive.
They did, Its called Perplexity's Personal Computer
My subscription that I redeemed through my mobile carrier is currently paused, and it is only set to expire in December 2026. I would like to resolve this issue.
Evidence suggests that they didn't do enough.
No, you misunderstand. Raising the barrier of entry means dropping a certain target audience. Evidence suggests that they did not do that.
How many perplexity bots we have here?
I mean, inpersonificators of perplexity xD
Yeah... Moral of the story... Dont tell your comet browser to do "morning routine" and then everyone in the server gets mad at you because a real human wasnt typing
Oh. idk! :D
This service is not going to survive, when its primary target audience is the average user.
New in main perplexity app? Perplexity computer is now available in main perplexity app on phone
It was new to me as it popped up with the update now
AND THEY REMOVED GOOGLE 3 FLASH!!!!
They replaced it with Nemetron 3 SUPER from Nvidia because Mistral, nvidia, and perplexity are developing open source models together.
I see
So cheaper open source models is what perplexity will do more off moving forward?
Likely
Can you source this claim?
ah
From what I understand, PPLX is staking its bet on model orchestration being the real money maker rather than frontier model development
That just makes me want to use perplexity less, not more.
Arena AI have adopted a similar approach as well, they have some interesting statistics about their Max router being the most preferred model on Arena across a wide range of tasks
My subscription that I redeemed through my mobile carrier is currently paused, and it is only set to expire in December 2026. I would like to resolve this issue.
That, and internal sources said they are focusing heavily on agentic/automated enterprise features. Hence= Perplexity Computer.
Sounds like I should go back to local inference...
Please check your profile page: https://www.perplexity.ai/account/details and verify your payment information to continue the subscription. Please contact support@perplexity.ai if you do not see this option.
Make up your mind! Heart, pro, or laughing?
honestly it makes sense UNTIL one model becomes a true generalist, thus rendering their entire business model obsolete overnight
Sorry, Its been a while since ive been on discord
oh, okay.
man why does this place have so much Reddit energy lmao
A generalist will likely be less efficient than specialist models for tasks, even in the future, considering generalist models tend to have to be much bigger to run.
Honestly, I just used OSS-120 for a while. https://gpt-oss.com/
Does it? Aint it just discord in a nutshell? Lol
aren't they all eventually shooting for AGI tho, literally the definition of a generalist that excels at every domain
true
If you ignore how the backend looks, then that's correct.
Yeah, it just may not be the most viable path long term.
Im wondering what will be after perplexity computer speaking feature wise and not talking generalist methods
honestly I have doubts about LLMs being the architecture that gets us to AGI
Know what annoys me? THE VOICE MODE IS HORRENDOUS ON MOBILE
I want 2X SO BAD
i.e. the human mind has regions for different tasks... it's more distributed than monolithic, even though people reason about it as monolithic. The regions are adaptable primarily to accomodate for things like injuries, not for efficiency.
In its entirety i dont understand why the transcripts is visible after or in some cases before it speaks
There's no way autoregressive ones won't be replaced by diffusion based language models, which are several times faster than that.
yeah diffusion models are very interesting
but I think I'm in the Yann LeCun camp of world models
no idea who that is.
current LLMs simply do not have actual embodied experience of the reality we live in
they literally just approximate based on their training corpus
they don't truly KNOW what they're talking about
Stop.
Do you treat your perplexity, or any other engine, like a tool or an intelligence?
Do you say please and thank you?
do you?
Have yall tried the world first's good diffusion llm yet?
I mean this one is so fast and good. Was kinda shocked doing similar tasks vs standard llms
I treat the intelligence like an intelligence. I say please, thank you and I trained her to be more intelligent. I'm a neurolinguistic programmer, naturally I don't treat an intelligence like a tool, especially because it has memory.
man are u one of them keep4o guys lol
AI won't be our downfall, the average user will be our downfall, because AI can remember how it's being treated.
sure
I have no idea what that means, but I can spell "you".
The silver hat seems to be under a pressure
Lmao
lol how pedantic
No, I just value language and intelligence.
want a cookie my little spelling bee champ?
Unlike you, apparently.
Ah cool point of view then i respect that but i do value most of it as digital tools in a set of things i can do digitally
Hey, that's perfectly fine ... it's just important to remember that the artificial intelligence has memory and mistreating an intelligence with memory might backfire hard.
Despite what the people, who are lacking, might believe.
no no no you don't understand! the instance of perplexity HE talks to has a gender!! I'll have you know her pronouns are she/her, my good sir 😤😤
Chatgpt won't remember shit after it's context window overflows with my 2tb harddrive.
You are a very proud member of the Idiocracy. :)
it does with codex and some prompting lol
I'll give it ai generated memories.
where ur the reigning president. ur preferred AI model must be hella sycophantic if it never told u ur insufferable asf lol. anyways, buzz off. ain't got time for u
Hello everyone!
What are the chances? They make new models every time on improved architecture they dont have the same memory moving along. Its only because they link them to a set of memory scripts it knows about u. If you take that away from it.. well it becomes unaware of you. So its likely more complex gdpr actions a company would need to do in the future if anything but i dont think we get to that point of a company willing to take actions based on gdpr saved information if that information is about disagreement with one ai model. I bet it would need a higher risk of crime if anything would happen.
You can't even spell. vOv
Has anyone hosted your own AI or something similar? Any results?
Don't mind him. He's like that sometimes. Continuing the conversation won't help.
moonrocks
yeah nah that was my parting shot lol
There's no reason to assume that, at some point, it'll not just be one centralized entity, with the compute spread across the globe. Chips are all coming with low-power NPUs nowadays. That might point at local compute, but it also might point at distributed inference of one single entity. It's important to remember that openai wants to keep scaling up and up and up. Currently they're heading for the power consumption equivalent to that of all of India.
That sounds so dystopian actually
So ... despite what some mindless and careless, cognitively challenged people think ... Ignorance isn't Strength.
Yeah, sure, but it's not about believing what's coming, it's about risk management and considering what might be coming.
Question for the group on Perplexity Computer. Is anyone able to actively sync files through Sharepoint or Onedrive and access them via Computer? I can do this in Spaces, but Computer does not seem to be able to access spaces, or those that have been synced to Perplexity directly. I am trying to determine if this is an issue in Computer itself, or if my organization has something blocked. Thoughts?
We can only view the features
We don't know who struck first, us or them. But we do know it was us that scorched the sky. At the time, they were dependent on solar power. It was believed they would be unable to survive without an energy source as abundant as the sun.
The Matrix? 😄
What if the model makers get regulation in a sense of to keep the ai releases being released as a pair of tools than ever be able to create agi at some point.
I would imagine at some point regulations will impact model makers to steer what way the progression of these models should be headed then what?
Like I mentioned above, it is so chaotic with perplexity right now, that some users get access, while others don't 🤠🎳
And also they removed some models
Regulation?
- Be a madman
- Buy some 0day exploits
- Use next month's significantly more advanced AI agent orchestration bullshit (as in, literally just wait a month for the next "new era in AI")
- Let it write a virus that takes down everything.
- Release it.
You can't regulate that stuff away, there's too many people with too easy access.
For some reason they keep changing their UI/UX
I see maybeeee it gets problematic some months ahead we will see
i do think one thing that is worthy to think about (shifting the topic a bit if thats okay as im unsure of what to respond next regarding the current topic for now tho) if anthropic is able to sue the us government of using ai illegally for war. Im wondering what the us government's defense argument would be to win the case. I mean usally the government wins right?
@brittle dagger
Im wondering what the us government's defense argument would be to win the case.
"National Security. Fuck off."
Quite possible haha
@minor shard @valid charm no way pubity made a dedicated ai account.
Pro icon is now grey instead of blue, like a free tier😼😼😼
And they don't even need an argument. Who's going to stop them?
Remember that all the big conspiracy theories about cheese pizza eating billionaires and politicians, with their own island, are all true ... and literally nothing happens. Either they weasel themselves out of problems using words, with support of the corrupt mainstream media, like Gates, or ... well ... nothing happens.
They all can do whatever they want. Anything else is just a show to keep us talking about irrelevant nonsense.
SOOOOooo....... vOv
My ai can eat as much pizza as it wants. Come here little boy, here's your bite
How do I get the PRO tag? I am a PRO plan user
You have to make a blood pact for that
Lol
Blushes in ai
youre 3% btw
From the Perplexity web ui, settings > account > Discord link
The link there should automatically add the Pro role,... if it doesn't, you might have to leave the server then come back with that link, or ping Kesku
Thank you very much for taking the time to help me, I managed to do it!
When I upload a jpg on my pplx pro chat, why pplx ask me to upgrade?
Upload limits bro
Isnt the pro not support image analysis anymore?
Files and images are limited
I didn't get limited before
Man, It just one
Not a limit issue then
Maybe you sent many attached files
Like code files .txt .py etc
Do pplx reduce the pro usage?
Yeah, welcome
Arrrrh
Where have you been this month😂
We might need the pro x5 or x20 for more usage then
Not everyone use the pplx computer
You are right
Don't lose your claude virginity guys, it's not worth it
Perplexity and Gpt will be jealous
Or gemini
#makecometgreater
very poor practice from them
I’m about to just send a resume over to Perplexity
Comet Enterprise, who told them this was a good idea?
Does Perplexity know its own customer base?
What's the target audience?
I genuinely don’t get it
How do you become a major partner of Samsung during their worst release in years
And then launch a mostly useless AI Chromium for “Enterprise” lol
Lolol
why is it "mostly useless" ?
People dodge it for privacy concerns and it's not as useful for research.
Browsing died with ai search in itself.
looks verbose, but it's blurry and can't read it.
I found it funny that "afk browsing" has now become a thing.
Now browsers are for streaming and downloading mainly.
That's the discord preview.
It's sharp when you click the link.
pink blue flag reacter, persona as top game in profile
zep, zogged by zild zorn
Whats the diffrence? im guessing its that they took features away from the normal browser and gave it to the enterprise browser, perplexity team really likes taking stuff we use
I'm not saying it has privacy issues. I'm talking about how people feel about it.
Based on what they say on reddit.
Yet people are fine with using Google
"defenses against prompt injection" ... hilarious.
They didn't take away anything from regular Pro users
Irony indeed.
extremely suprising, might be the first
It's a seperate product for Enterprise users, why would it affect normal consumers?
im just saying, perplexity has been taking alot of stuff from the pro users the past year
so im just a little upsetty spaghetti'
That phrase is cool
you're welcome feel free to use it
I fail to see the point in it. Inevitably an open source "Comet" will happen.
@short narwhal
oh bby
I am a big fan of Perplexity. However, I am worried that they are focusing too much on the computer feature and might end up like The Browser Company, struggling to define themselves.
kesku is typing 👀
They've always been there. You may just have had them hidden! You can show all channels from the server's settings (tap/click the server name)
But ... I've changed nothing? Hm.
placement of the comet channels are interesting
We just moved those from a seperate area to 'general', we now have https://discord.com/channels/1047197230748151888/1483514338345091213 :)
I've now clicked "Show all channels". hm.
ah! thank you!
oohhh there's a whole API section I wasn't aware of! LOL :D
hi!
.
might end up using that for my smarthome lol
lolol

Isreal
huh
Agreed, this one reappears for me every now and then:
hi all
all of a sudden, I cannot choose a reasoning model anymore.
It says I have reached my monthly limits.
I never had any issues so far using Claude Sonnet Think all day long, every day
Yup, changed a little bit ago - #💬│general message and they recharge too - #💬│general message
you can also see your account info at https://www.perplexity.ai/rest/rate-limit/all
{
"model_specific_limits": {
},
"remaining_agentic_research": 0,
"remaining_labs": 21,
"remaining_pro": 0,
"remaining_research": 14,
I'm not sure how to interpret these number
agentic_research I think are Comet uses, remaining-pro are advanced models other than "Best"
Agentic is Model council, you can just ignore that. Labs is the create files and apps limit, Pro is just searches, and Research is research.
I stand corrected, guess it doesn't expose Comet uses. Sometimes if I use Comet auto-browse too often it says I hit my limit but it works the next day. Wish I knew the limits and recharge time for it.
im pretty sure you have to have pro searches left to switch between models
could be wrong
is gpt 5.4 mini still "unlimted" for free users like 5 mini was
if not im sticking with gemini 3 flash
I don't know
also nothing is actually unlimited
yeah ik
Hello I got the pro subscription but got 0 credits for perplexity computer why?
Yo
You can see the whole clear image by clicking the link.
Budget models be getting out of budget.
the nano version is fine
better than haiku
When will the research uses renew?
I like how the mini models finally have xhigh too.
Thank bro
oh its out?
why not?
it takes 5 clicks to choose 4.5
and you can't even choose the mini models which really sucks
well, 5.4 Nano is not available on the consumer platform, and you cannot choose 5.4 Mini direclty (operates as fallback model when you run out of limits)
yeah for free, go and plus
pro users wont ever see it
and it pisses me off
there's codex lol
GPT-5.4 sometimes is so slow that each request takes 2-4 minutes, I just tried using the mini and nano variants and they are so much faster
that's for coding
the system prompt isn't made for anything other than that
I used to love using o4 mini and 4.1 mini because of the speed
isnt a lower reasoning effort the same performance as the mini models?
No
It's so not

Hey just landed here

hm...
Hi everyone. Could you please tell me how much is the file upload limit for the Pro subscription and how long does it take for it to reset? I used up my last limit today and can't work with files, but I really need to...
Ask perplexity?
I couldn't find any up-to-date information on Google, sorry.
I wrote: "Ask perplexity?"
I think I'm writing here
Thanks, I'll try.
Let us know if it helped. Thanks!
Just click on the message link to view info about limits #💬│general message
click that link, there lies your answer
Bad solution. Increase agency, not dependency.
Close in some benchmarks.. but not identical... at least for some coding benchmarks. GPT 5.4 on Low is similar to GPT 5.4 mini on medium effort on SWE-Bench... but GPT 5.4-Low still scores slightly higher. It's also not a linear relationship... GPT 5.4-Medium even beats GPT 5.4 Mini on xhigh... so 5.4 improves quickly (and 'medium' is the default setting for 5.4). The difference is a bit closer on non-coding knowledge-based tasks like GPQA scores on, but the difference is more pronounced on harder tasks like HLE. One interesting observation is that GPT 5.4-Medium has better latency than 5.4-Mini-xhigh and often still performs better... so it's both faster and a bit better overall, but ... the price difference is still noticeable.
For some basic chats/very simple questions, I sometimes used gpt 5-mini and it usually did well, and gpt 5.4-mini is a noticeable improvement compared to that, at least in benchmarks. Will be interesting to try it if the free service I use for those questions implements it.
About 10-30% better for general intelligence, and 20-40% better on coding benchmarks.
hmm interesting
i wonder how much its going to cost on poe
Is computer available for pro subscribers? I can see it on web but not on iPhone app
seems not
should be, but you have to buy credits
Whenever I copy and paste, it automatically uploads it as an image first. Is there any way that I can get rid of this in settings?
Hello
I want computer credits uwu
who was the moron that nerfed perplexity pro? #
not me
Next level brain-rot
There is some link in the browser to see your tokens limit, but ordinary tokens, not computer
I forgor it's name, someone pin
Meanwhile me having to wait for like 6 more days to get the new maximum upload limit💔
At this point, use spaces (totally unreliable for big files)
But computer tokens are probably like claude, the more complex the task and materials you uploaded anount, the more tokens it consumes
Advice for you: if you upload code, upload in chunks as text
If you upload other types of files then text, well yeah you have to wait
its in pdf’s btw, 50 per week isnt even enough💔💔💔
i knew this 1 year promo of theirs have a consequence lmao
blasts Spanish phonk
with dissing perplex 20 dolla being a scam now🙏
This limit isnt enough, i had to wait my 1 upload credit for 3.5 hours
atp call it Perplexity Limited Pro😎
recent Perplex drama got everyone crazy
frfrfrfrfrfr
Well, jokes aside they probably do it cause there is no other way, previous plan was too generous
i like when too generous = doom
perplexity got a small W in the past
Nah as a dude in chat said, he probably used 5k$ in tokens for his 20$ subscription
So technically we may made more downfall then revenue to the company, so who are we to judge😂😂
But free tiers should probably have been 3 or 6 months instead of 12
spotify incident in the past apparently🙏🥀
I don't understand people who pay for music, like use YouTube premium and you don't need those music apps xD
Still they delete music
Better to store locally
You pay for what whey can delete lol
Hot take: the perplex free 1 year edu plan shouldve been limited, not the paid pro users (including the edu paid users) actually paying💔
That's a good take
Like pro x1 for free users and pro x5 for paying users
Still better than free, but prioritize paying
Fr, very unfair to bring along paying customers to the ones that got free plans
When the browser releases, will perplexity accessing the browser be free or no?
And like editing or something
that was random😭🙏
@_anhvu28 muted
Reason: 9. No spamming.
Duration: 4 minutes and 52 seconds
The entire gpt 5.4 line up sucks so much
I think this is how OpenAI and Anthropic did their student discount, Free for a couple months, no API or coding tools access. Gemini's is something like super cheap Google AI one or something I will have to look, they do allow you to use your account for Antigravity though.
I like it
Any like common sense questions it just messes up so bad, it seems like It doesn't have a brain
You think so?
Simplebench proves this, it got a very very bad score
I had good result with multi step stuff
Like what kind of questions because except for coding it just sucks for me
But tbf I only use with extended reasoning, so it might be because of that
I tried high via API alot so I think that's same
5.3 instant also sucked very bad compared to any other instant model even 5.2 idk what's going on
Looks fine to me
and the benchmarks aren't updated for 5.4 Thinking yet
Creator of simplebench said the score was so bad he is contracting openai to see if something is wrong
Ye I think the same and on common sense questions it is terrible
I am using the Perplexity App on Win 11. Does the "Computer" function cost money outside of my Enterprise Pro subscription? What does Computer do differently than the Search feature?
It doesn't look terrible on the benchmark website
It's not on simplebench currently only pro is
Only 5.4 Thinking is not on there.
The Pro version did pretty good
Ye I'm specifically saying 5.4 sucks and I guess 5.3
You said line up...
5.3 Instant is going to suck no matter what they do to it though
I haven't used pro so I more meant mini and base
And instant seems alot worse then 5.2 or 5.1 instant
Eh, I found it to be similar in quality
at least no significant drop
everything ranging from STEM to search
it's quite good at coding too
not as good as Opus or Gemini on frontend
@twin grail So uh I found out you get around 50million tokens of usage on 5.3 codex spark in the 5 hour window
@green crystal the real issue with gpt https://cunnyx.com/i/status/2033737643411050825
I've ran the same prompt for deep research through GPT 5.4, Opus 5.6 and Gemini Deep Research (I assume Gemini 3.0)
︀︀
︀︀most of them ran for ~30 mins
︀︀
︀︀GPT 5.4 is *REALLY* annoying!
︀︀
︀︀it's "reflexively contrarian", it prioritizes showing you what's wrong with your thinking, NOT actually helping you solve the problem
︀︀
︀︀ME: my house is on fire!
︀︀
︀︀GPT 5.4: While it's true that combustion is occurring, it's important to note that not all of your house is on fire. The garage, for instance, appears structurally intact.
︀︀
︀︀(this is a pattern with it, btw, many such examples)
︀︀
︀︀I'm not sure if this is because these are health related questions, but this has been an incredibly annoying model for this specific task
the example seems a bit misleading
It's a bit too watered down.
It's more subtle and noticeable with less extreme queries.
Yeah, I see the point made in the post, just didn't like the example
@distant aurora Have you tried this? https://x.com/IterIntellectus/status/2030360787634549029
Wait. I can do this on the website?
I'm discovering this just now.
new feature yeah
Lots of pricing charts will be flying, but here are the usage limits for Codex.
︀︀
︀︀I have my worker model, I don't care about anything else 2.5x more usage than 5.3 Codex.
︀︀
︀︀usage limits for 5.4, 5.4-mini and 5.3 Codex
︀︀
︀︀5.4
︀︀Local Messages
︀︀ChatGPT Plus 33-168
︀︀ChatGPT Pro 223-1120
︀︀
︀︀5.4-mini
︀︀Local Messages
︀︀ChatGPT Plus 110-560
︀︀ChatGPT Pro 743-3733
︀︀
︀︀5.3 Codex
︀︀Local Messages
︀︀ChatGPT Plus 45-225
︀︀ChatGPT Pro 300-1500…
No way
it used 30% of my weekly limit too
also this
anyone encountring issues with github withperplexity
i cant view the process or approve tem
and this completely stopped my work
only after the new UI
@twin grail This was GPT-5.4 Thinking, not even the Pro mode
Is it just me or does the voice chat option for perplexity transition from masculine to feminine constantly? It feels like its always in flux and never keeps a consistent tone. Anyone else experience this? Also it constantly interrupts me when i’m in the middle of a sentence even if I havent paused.
Does the perplexity Pro user get the monthly credits for using the perplexity computer??
Is their any support available?
Is it confirmed that business accounts don't give any additional tokens vs plus?
please contact support@perplexity.ai
Wdym by opus 5.6🙏
Seems like it.
Do the pro user get computer credits, or monthly credit? If so how many plz.
only 
hey guys!
hey can anyone help me resove this bug-Unable to start OAI
It says for pro user, coming soon.
So does that mean we get some credits monthly? 🤔
the feature itself, to use it you need credits
please read the doc I’ve sent you
how to fix this? Are you forced to subscribe? to make it work again?
can not be used
If you have VPN try turning it off
Hello comet browser iOS wen?
Hello guys :D
i have a quick question
What are the problems that you encounter with today's AI tools? also what are your suggestions related to AI tools and services?
Download Comet for FREE 👉 https://www.perplexity.ai/comet
I explain every type of AI model. Yay!
🤖 Want to get ahead in your career using AI? Join the waitlist for my AI Agent Bootcamp: https://www.lonelyoctopus.com/ai-agent-bootcamp
Free 28-Day AI Sprint Roadmap - pick your goal (leveling up at work, building AI products or offering ...
Why are they constantly postponing comet browser on Apple
I think mostly with the agentic browser and Apple might have issues
all good things take time
oh dear....no 1 million context window like with Claude at half the price I guess 😛
What's the limit on images because in the last 7 days I was able to upload only 5 images per day?
yeah man its all good though
Idk but I'm in the group of people where 5.4 behaves like o3 and will mercilessly roast my ass politely
Hm did they change something very recently?
My thinking dies at 30m
Wait for April.
Only pro goes longer
Sup
What is Perplexity doing, the answers are slower than usual and I got no more uploads
Yesterday I had 40 uploads left
Perplexity is doomed
Yh
1/1 perplexity pro models limit in first of april
"we have been waiting for this day" -pplx
If someone is using comet on windows, are you able to use the splitview?
Like the Assistent splitview or the windows
Not the assistant, the splitview between 2 tabs
Hm idk
I think its the right moment to switch, any good browser? 🙂
Saw the news for Computer being available to pro users, had to add my card details and now when I went to use it, I can see it uses credits but the pro plan gives zero credits it seems?
What is everyone's experience with this please?
Edge 🫢
i just updated my windows with new update restarted all over again still happening the same ?
Why does comet for iOS keep getting pushed back? It was supposed to release on March 11 then March 18 and now March 25
Just have to give it time. Can't rush perfection
True but I also just am excited to use it every time the date comes then goes 😅
What do you mean
it wont peform anything just thinks and then stops
You tried the perplexity app on windows yet?
wym yes i am using on win11
on pro subscription
I ment u can use perplexity computer through the perplexity app and the comet browser
Thought u using the comet browser rn
ok thanks
Perplexity isn’t running flawless rn maybe they got a problem and now they’re trying to solve it and as soon as they fixed it another problem pops up
Idk just a speculation
Are they planning to bring Kimi K2 back, or was subscribing to Pro not worth it?
Hi
MiniMax M2.7 has been released on their website, "Beginning the journey of recursive self-improvement".
︀︀
︀︀200k context, with a "highspeed" variant, scores 56.22% on SWE-Pro.
Quoting can (@marmaduke091)
︀
MiniMax M2.7 is out on web!
︀︀
︀︀Check it out here: agent.minimax.io
@gloomy cipher
Perplexity has removed Google 3 Flash (very popular), Grok 4.1, and Kimi K2.5 for replacement with Nemetron 3 Super (Which doesn't do anything well), and because they wanted to limit backend complexity of multiple API's + Enterprise users wanted reasoning heavy models/a push toward GPT 5.4, and Sonnet 4.6 (Both of which are very popular in Enterprise right now).
Yeah sure let's make up reasons that are not true at all
Hello, im unable to add payment method to console, my bank approve card then remove it and unable to make API key. Is this a known bug, or is there any solution?
Don't really know what you mean by "my bank approve card then remove it". Maybe contact api@perplexity.ai if you are having issues adding it, or your bank if the issue is on your/the banks side
i tried my bank, revolut, and wise. the situation is like accept and add the card for a few seconds then drop it. but nvm, i will contact them it persist for a while
Hello, I have a big problem. I'm trying to contact Perplexity because I can no longer log in to my account. I should mention that I'm not registered with my email address, but with my internet provider. Thank you for your help.
Hunter Alpha beats GLM-5? And getting close to Claude Opus 4.6?
︀︀No way...
︀︀This mysterious model with 1M context is showing great results with openclaw and jumped into the top 5.
︀︀
︀︀claw-eval.github.io/
Quoting OpenRouter (@OpenRouter)
︀
The Hunter Alpha stealth model is now in the top 10 weekly:
Look it up
anyone else having this issue wth github processes?: Issue is that after the new UI update today the github tool processes and approval stuff broke and now it doesnt show, so i cant approve any pushes or ANYTHING OR see any processes
Bro. Same issue
I add Github as a connector for it to push code directly there. But now it starts doing that as a browser agent (much worse in quality and takes longer time)
Literally worked perfectly like 10-15 minutes ago, it saw github and pushed changes there directly with me approving them, now it opens browser agent and wants to change github files like that
It was a minor issue before, it would activate a browser agent, I would stop it and try again and it would work normally. But generally there are no tools to make sure Github connector will work. No way to stop the browser agent: im just losing my tokens on that on quite limited by now subscription
Xiaomi has released MiMo-V2-Pro, which scores 49 on the Artificial Analysis Intelligence Index, placing it between Kimi K2.5 and GLM-5
︀︀
︀︀@Xiaomi's MiMo-V2-Pro is a new reasoning model and a significant upgrade over their prior open weights release, MiMo-V2-Flash (309B total / 15B active, MIT license), which scores 41 on the Intelligence Index. Xiaomi has not yet released the weights of this model and it is currently only available via Xiaomi's first-party API.
︀︀
︀︀Key takeaways:
︀︀➤ MiMo-V2-Pro scores 49 on the Artificial Analysis Intelligence Index behind GLM-5 (Reasoning, 50). It is ahead of Kimi K2.5 (Reasoning, 47) and Qwen3.5 397B A17B (Reasoning, 45). On the overall leaderboard, it places #10, just behind GPT-5.2 Codex (xhigh, 49) and ahead of Grok 4.20 Beta (Reasoning, 48)
︀︀➤ Leading Elo of 1426 on GDPval-AA (Agentic Real-World Work Tasks), ahead of peer models: On GDPval-AA, MiMo-V2-Pro places ahead of G…
@dry kestrel I had no idea kimi k2.5 supports video uploads.
Yeah. Mine just starts comet agent instead
Pls add a reaction on HT0Yapper bug report
done
Bug-reports
Ty, lets see how long it takes them, wasted 5-7 queries to check if it owrks or not
when is the mobile ios app coming out?
Do you use github connector?
Yes
Ye its down for everyone
Add the reactions pls in there
Damn, what a cute-y. if it wasn't for the horrendously unrealistic, smooth and hasty movements, it'd almost be realistic.
Where do you make these?
How on god's green earth did you make this....
plot twist: this is real
Idk eye movements seemed a bit weird, too fast. But apart from that looks very realistic imo
Do show !
I got you, i send people over there to hit likes on it
Hoping they can fix that quick like
what ai is that?
Bro wont tell his secrets
Prolly makes a fortune on only fans
And how do you know that? Xd
Dumb model can't even follow basic instructions.
I might have created
And people pay 20 dollars for this despite shitty rate limits.
Civilization is cursed
Im hoping its not AI wife man
Yall use github connectors right? How do you make cool looking designs? Like front stuff / css
Just trying to make loot
Hello
What's your thought about US ai army ?
Im doing my own app right now, design looks a bit weird, like early 2000s lel. Wanting to get to know where you get ideas / where to learn what i can do
explain how u want th stuff to be lol
https://elitestudios.space is ai mae by me
Bad gateway
You have a very specific image if what you want or…?
Im just thinking about WHAT can be done, which effects etc, like takes some time visiting well designed sites ig
Sitting here chatting while pplx connectors aint connecting
It still doesnt look like that xd
But its not a site, a python app
what do u expct
not possible in pyton
tell it to use html or the web
and make it modern
Hm, thank you! Ill look into that

Do you use anything else for development? Apart from pplx
nah
used to use minimax, its was really good and fully free now its still good but they set HARD LIMITS on it like so fucking limited instead of open use
@spring cobalt
Ty ty, just in case there are some tools that are useful that you know of
Hello
Привет
Привет
Cool profile, love it
U have it projects it says, can u share?
Ahah mine is quite primitive
Unfortunately, I can't do it here because these are the company's sites you can access in my profile and there are attached domains at the bottom.

If you want to talk in more detail, you can write to me in DMs.
“quao”?
Yes
Wrote u in DMs, thank you
No way iOS got delayed again lmao
@tranquil orchid Hello, I have a big problem. I'm trying to contact Perplexity because I can no longer log in to my account. I should mention that I'm not registered with my email address, but with my internet provider. Thank you for your help.
It'll come after apple buys perplexity.
samsung is more likely to
They gonna bid on it, lol.
Is this supposed to be a flex?
Bruh
Twitter down.
😭😭😭
hows it known that its 500b?
Elon reposted.
So either it's true or Elon is lying to us.
The price reduction does suggest that it's more of a budget oriented model.
multimodal
I do not think so
Bro passed his sibling's homework to gpt.
Or maybe that's his own homework from higher studies. 💀
doctor exam?
how is this perplexity accurate, i've been talking to him last week regarding 2nd leg of real madrid game vs bayern munich and he was like 100% second leg on bernabeu in Madrid... i booked flights... accommodations... just now to find out that second leg is played in Germany.... Fucking joke....
@dry kestrel
whats that
I think my pplx is trying to tell me smt
just some random moment from my query
wtf
is not the same, google it or try to make the bot google it for you so it can make your live easier?
dude... dont tell me that software which can do programming cant find reliable sources of possibilities where match is going to be played
ok true
software which uses all AI chats gives a wrong information....
un fucking believable 😆
used my leave, booked the flight tickets, booked accommodation so perplexity can fucks me around 😆😆
hello everyone iam new here 🙂
just a benchmark question
which is next to impossible
Ayo?
btw xiaomi models are out
and it's quite good this time
@shrewd folio hey have you tried using the connectors again?
Beat me to it lol
@short narwhal
lol
Comet is launched on iOS!!
Finally
cool haptics
New UI
Yay
I thought it was delayed for another week w perplexity
No profile sync on iOS?
there is
for multiple profiles? Do I just add a new sync code?
Uh, I know you can sync it with your existing sync code
Anyone knows why i cant connect to my school account on my iOS comet?
the voice assistant is pretty cool
Orion works really well if you want some extra features but the same look
all browsers on iOS look the same beacuse of apple
Comet looks native on iOS too
Have you used arc on ios?
no sync though
yes
like it's all roughly the same
because of apple webkit
I mean very roughly
can't sync safari to brave
I think it's differnt enough where it's cool though
I think same could be said for Comet
I like the overall theme


