#💬│general
1 messages · Page 567 of 1
I mean. Right now since early this morning. I now have -59000 computer credit. And. 75 %of my prompt wont work.
U gotta agree that there wasn't a scandal about Comet being hjacked in a real case scenario because most of them were patched
how is that even possible
u can't have 59k computer credit btw...
i have differnet issue
my responses cut off mid
its frustrating
see
Did u
did what
been chatting with the AI for a while ?
yes i believe
Deep
i ususaly change models to fix this
oh ye that's why
but i have to do this twice and same shit happens again
It doesn't fix it
never had that issue with other ai
U just gotta need to open a new chat
Because if they does, they will start to hallucinate
all context will be lost :(
yea i get it
You can still recall it's memory
instead of flooding it
with tons of stuffs
Perplexity having the context windows i think is above 192k+
the same with chatgpt and claude
no clue
192k+ of what
context windows...
whats a context window
U should do some research
the ability of how much information that a LLM model can handle
Howcome 10K credit cap for pro users any way to get around?
before hallucinating
that i get it but what about the unit? like say what does 1 context window mean
okay so
yes
yes
yea 192k in that case is sufficient
reasoning
No clue what's the real context window
guess i been just abusing my perplexity too much 😭
it just the lowest context windows possible
could be higher
If u want to do long context coding
go for gemini
Any-body?
1M context window
should help
I usually parse my document with gemini
then go for perplexity when i do research
ye i do too
just a year ago, i used to copy paste code from online genai
now even the ai can write code and run stuff for u
kinda scary also
but heres the thing. only in my class managed how to utilise it utmost
my github co pilot credit is drained
used it a lot
i am gonna put student pack to get that for free
its useful
Is this on perplexity. Where do you acces it
i don't use copilot...
even tho i have copilot teacher
its in built on visual studio code
It just doesn't meet my requirement
plus i use co pilot for general stuff
yop
i use codex for it
like discussing some topic if im bored
i heard its expensive asf
Nope it's on Google Antigravity
like someone said 30$ per day??
haha not really
No
maybe i misheard
the same as google ai pro plan
20$ for plus plan
u have access to codex
i use ChatGPT Pro
So the limit is way much higher
i am gonna put some money onto one ai subscription
its a life saver to pump me projects and stuff
U should invest on codex
ngl
Antigravity pro rn
is trashed
yea been thinking
i still have credits for anti gravity i believe
They nerfed it
and i thoguht i been putting unnecessary images when i could just copy paste text
U can tell by me burning 700M token on codex in 4 hours
its good while it lasted 😔

cooked
nerfed hard
I really
considering
switching to claude max
still none of my collegemates know this piece of software
they dont even know they have ai agent on their VS code
that's ur advantage
never tried claude
good quality
i introduced it to some of my friends and i still have to prompt for them 😭
ill keep that in mind
business ?
im compsci
or u bought the plus version
i got pro version at a discount
instead of 20$ it was 10$ per month
for 3 months
that's not pro...
i want that offer again
still it was useful
unlimited image and file uploading
buy for u
how
how much
not unlimited
u should
read it again
dm me ur email
only limit thing i faced is his ultra deep thinking feature
i'll invite u to my chatgpt business
can i trust u
dw im playing
redemption arc
do not redeem
that's...
i had to do it because Antigravity being nerfed so much
ah
sure
lmaoo
Reason why perolexitynis failling right now
Perplexity isn't failing right now?
should be 500 per month
and 2k bonus
for the first month
go to computer
u should see a battery
It's more than what OpenAI gives
A lot of user are reporting issue. Including me.
You need to be more specific.
Yes but it's way worse
I pay them $200 a month and their agentic capability is worse than on the $20 Pro plan on Perplexity
Never
touched
OpenAI agentic feature
even tho i'm on pro
i prefer perplexity way more
the 20$ plan
totally ditched OpenAI out of the play
u should buy max plan atp
max is kinda good
Lot of error message that still use credit. Negative credit. (Im at -59000) but no trace of what use those. And instance. That delete parts of the already completed work
u can't sync them all at once + the quality is way behind max
Same boat. Same problem. I wonder if anyone from the Perplexity actually checks these chats.
and be specific and provide screenshots and examples of the issue
You can see the total amount of credits used in a thread
Already done it. Its where I saw that im not the only one having those issue. Most of them start less than 24h ago for allmost all of those
Hmm, I've had a Computer task run 3 hours ago and it was fine. Perhaps also email support@perplexity.ai
No
The comet agent ≠ pplx computer
2 seperate things, the agent in your browser does not consume computer credits
It's a rolling window, slowly regenerates #💬│general message
I don't work for Perplexity...
Computer was also working fine until it show -59709 credit lol. Concil is the worst right now for me. 30% fail.rate (stuck or error message)
hm, I've only had council time out on very very long threads
OK, Ill not make any more jokes then
anyway, perplexity is stuck in thinking and I have no clue what is going on
just tried different browser doesnt work
how about on a different device?
ok I got something
hmm interesting
ok it works on my phone for unknown reason
but not on laptop
someone at ANTHROPIC just showed CLAUDE finding ZERO DAY vulnerabilities in a live conference demo
︀︀
︀︀claude has found zero day in Ghost, 50,000 stars on github, never had a critical security vulnerability in its entire, history...
︀︀
︀︀it found the blind SQL injection in 90 minutes, stole the admin api key, then did the exact, same thing to the linux kernel
@green crystal no instructions, no rules, no goals. Like what is the scoring even based on 😭
There are technically goals... it's just that the game rules/goals aren't 'provided' up-front. So if you're going in blind, you have to observe, orient, decide, then act on new information depending on what you try. Some of the goals dynamically present themselves based on those observations, and then progressively iterate on difficulty based on those same goals through different rounds (i.e. matching colors in round 1, but several of the targets may be the same color in round 3, so you have to figure out based on the shape of what you're moving, etc)
It would still be possible to provide prompts with the rules/goals manually, but to 'win' the context to the AI should just be the puzzle without that context - because humans playing blind also eventually figured them out. (Though I highly doubt humans got 100% on their first try... some of the puzzles are actually quite difficult)
I also don't think they gave these puzzles to random people on the street, if they did a high number of them would've just given up, so that 100% number is very misleading. The 100% is probably puzzle enthusiasts, and still given several rounds.
If you give all the ARC-AGI-3 puzzles to a room of random people, that number is going to be much lower than 100% guaranteed.
You can play the games here (scroll down to the select a game section): https://arcprize.org/arc-agi/3
Something on that laptop, like antivirus or VPN or a security setting.
dunno, everything worked yesterday?
Things can get auto updated to a new version.
Im out of ideas what could it be
Maybe try laptop on a different WiFi network?
Does perplexity pro have a monthly usage limit for comet automation?
I’ve never had it come up before
hi i can no longer use other APIs like claude and gpt, everytime i try to use them they changed it to 'model' automatically i have perplexity pro
anyone face the same problem?
did i used up the pro search credits or smth?
Hey all, whenever I try perplexity (from yesterday till now), I keep getting the “unable to start thread” message. Tried to log out then login in again, restarting, even using a different browser. Is there an ongoing issue or not? My subscription is active and all should be well
Hi, i've a question, on discord, telegram and whatsap is it deprecated the bot?
guys anyone knows browser which has as crisp ui as opera noen like honstly
see no rounded edges nothing
yet so minimal and clean looking
but its paid 🙁
#💅| bitchin-n-complainin
d
😌🤣😌😌💔😦🫧🥵😱🫧
sup
I want to do an internship, paid or unpaid, but it should be legit, not any type of scam, because my college demands an internship certificate in the 2nd year. Can anybody help me or guide me on where and how to find a real valuable internship around 2 - 3 months? I want to really do some valuable work for industry and gain practical knowledge.
Can perplexity work inside code editor like VS code? Like Claude codes create and modify complete code base, can Perplexity does that?
Hi my friend Joe Rogan ... He know I like Aaron Travis kk?
Got it understood permarketers
?
Kevin fallon 9498
But ... Gosh I do have that name change form filled out in the spinning dolly stand
Fine I'll Wynn
Futurr . The prediction market
I taught pond5 how FanDuel cracked gambling . Bang
Good morning 🌞 laggard
s
WuvUlots
Hi
roblox exploiter imagine
What
The computer toggle is a mess. Why queries written here are rerouted toward computer (and failed because of lack of credit) with no way to deactivate it ?
Is there a way to get perplexity pro for one month currently. (Using student mail or reffferals)
@strickenx. muted
Reason: 9. No spamming.
Duration: 4 minutes and 51 seconds
Hi there, I'm having problems with perplexity sincer 4 days. When I write just say that the thread cannot start....I'm Pro user........that happened with Chrome browser, with Edge went well since today.....Using perplexity on my smartphone goes well
what could that be?
please report here: https://discord.com/channels/1047197230748151888/1487828702556459018
thank you!
also try to resolve with this: #💬│general message
thank you!
I wrote here everything I tried....that with the extensions already tried
Hey guys!
I'm having some trouble buying API credits. Every time I try to top up, the transaction gets stuck on "Pending purchase" and never goes through.
I’ve already tried:
Multiple different credit cards.
Two different browsers and my mobile device.
And my OKX card, but that one was declined instantly (it didn't even try to load). Does anyone know if OKX is blocked or not supported?
Has anyone else dealt with this "Pending" loop?
Bruh. Google Ai studio is horrible.
First it makes my messages in the thread disappear.
Then some of my messages keep coming back after refreshing despite deleting them.
please contact api@perplexity.ai
what is the Perplexity discord AI tag?
how is perplexity interjecting itself into all of my hardware and software?
Has Perplexity still not specified in which regions image generation is and isn't available?
They have your soul
Hey, has anyone seen the pitch prompt for the stock competition yet?
Should have been released today
Ohhh thanks! i thought all pro subs got the credits. btw what is that "computer" for?
it's not?
Represent your server with Server Tags - 4-character identifiers that members can display proudly on their profiles, anywhere.
What this article covers:
What are Server Tags?
How to Use Server Bo...
Btw, now that you talk about tags
Hoping for some help with the Merchant Program!
I submitted the application for our brands and followed up via support@perplexity.ai, but haven't heard back yet.
Does anyone have ideas on who else to reach out to? Thanks in advance!
cc: @vague pagoda @halcyon coral @wraith crow
How to get free Perplexity Pro for Kazakhstan citizens?
Pplx app didn't know about that
it was super funny for me because I was signed up for perplexity pro offer and haven't used it for several months i think it's about to end it's because I kept signing up on chatgpt because it didn't create images like chatgpt then just today found out you can choose the image generation model from the all settings
why the confusion?
Heya, I signed up to the student perplexity AI one year trial in October - it still has a couple months until it expires but my prompts are no longer working for "insuffient credits". I've barely used any of my credits for this month as I've been using claude, so this has struck me as strange. Have they suddenly changed their terms and conditions or is there a way for me to fix this?
is comet down?>?
📣 Lessons from building voice agents at scale
@perplexity_ai breaks down how running voice with the Realtime API in production shaped their approach to context, audio pipelines, and turn-taking in real-world environments.
@short narwhal
Yo
I emailed Perplexity help-desk and managed to get a response. Would recommend the same! No longer an issue here, thank you 🙂
whats wrong with it
bug
fixed using VPN
that can do it

Hey guys, is anyone having trouble generating images within Perplexity? I was only able to generate two images with Gemini 3.1 Pro, and then it says the server is having problems.
Hey when is the pitch prompt being released for the Computer Competition?
@vague pagoda
perplixity is racist , it needs a vpn
Qwen released the Qwen3.5-Omni model family, designed for native understanding of text, images, audio, and video.
︀︀
︀︀- Qwen3.5-Omni-Lite
︀︀- Qwen3.5-Omni-Flash
︀︀- Qwen3.5-Omni-Plus
︀︀- Qwen3.5-Omni-Plus-Realtime
Citându-l pe Qwen (@Alibaba_Qwen)
︀
🚀 Qwen3.5-Omni is here! Scaling up to a native omni-modal AGI.
︀︀
︀︀Meet the next generation of Qwen, designed for native text, image, audio, and video understanding, with major advances in both intelligence and real-time interaction.
︀︀
︀︀A standout feature: 'Audio-Visual Vibe Coding'. Describe your vision to the camera, and Qwen3.5-Omni-Plus instantly builds a functional website or game for you.
︀︀
︀︀Offline Highlights:
︀︀🎬 Script-Level Captioning: Generate detailed video scripts with timestamps, scene cuts & speaker mapping.
︀︀🏆 SOTA Performance: Outperform Gemini-3.1 Pro in audio and matches its audio-visual understanding.
…
Man I love qwen so much.
The Most Generous Chinese Company
I hope they always leave a free tier with a usable reasoning model.
Is a Shame Perplexity now doesn't have even a Single Chinese Model
Is Inevitable
They can paywall the top variants if they want.
China is not trying to Make Money
They just want to Win Ai Race
Chinese Companies are Subsidized by Chinese Government
China has Cheaper Compute and Labor
@salty101 muted
Reason: 9. No spamming.
Duration: 4 minutes and 53 seconds
China Want The World to Rely on Chinese Ai
Business comes from Businesses and Enterprises
Free Users are Organic Marketing and Training Data Collection
More then Perplexity?
Quality wise, maybe.
Tbh Ai Studio has always been Buggy
yeah so sad such a small company is struggling with compute /s
When China Wins Ai Race
I wonder where all that money and stolen data is going.
It will be One of The Happiet Day of My Life
So much data yet the model is so shit.
The Models are Just Benchmark Gamed
Make News Headlines
Stocks Go Up
Investors Satisfied
Gpt is smart though.
Claude and GPT is Best Duo for Now
Google and grok benchmax. Grok gave up for now.
Exactly
I have mixed feelings about Deepseek.
It's numbers do seem benchmaxed compared to its actual performance.
Still isn't as dumb as google or grok though.
Deepseek Owner Company is Ultimately A Hedge Fund
Deepseek is a Side Project
Eventually Should Improve
Eventually Gemini could be Usable in Real World Use Cases as Well
I wonder if eventually we will see progress taper and models become hyper specific for certain use cases rather than a fight for general useability. Gemini is my preference for research/shopping assistant, claude for writing, etc. but I wonder if it will become even more diversified
hell no
absolutely no
That is my Prediction Because A Generalist Model can't Perform Best on Everything
While a Specific Model can Perform Almost Max
for their benchmaxxed versions lol
the goal for agi so far is for it to be a single model
and they're heading there sort of
next gen models are being hyped up so we'll see soon
@ader1472 muted
Reason: 9. No spamming.
Duration: 4 minutes and 54 seconds
so if anyone is interested
perplexity works on my other laptop
but cannot get it to work at the first one
I hope we get some big breakthroughs soon. Working on an animated movie right now so I'm waiting for google Veo 4 and the next grok imagine release
You can use Seedance 2 for Now
Not available in the U.S. right?
I tried the vpn route and had the issue of needing a localized phone number. Maybe with their latest country expansion I'll have more luck
HI
nOT if your Elon and can fund TerraFab and build space based AI data centers.
@maiden lion wassup bbg
I’m a single-user “power” user of Perplexity’s computer product and I burned through the max plan plus bonus credits quickly. At $200 for 10,000 credits, the refill cost is unsustainable for my workflow.
Does anyone know:
-
Whether Perplexity offers (or plans to offer) a mid-tier or alternative pricing option for high-usage individual users that is below enterprise pricing?
-
Any existing pricing structures, discounts, or negotiation paths for solo power users?
-
Practical workarounds, cost-saving strategies, or alternative tools that provide similar functionality at a better price for heavy individual use?
-
Other users’ experiences or barriers they’ve faced continuing to use Perplexity for high-volume workflows?
Context: I’m not looking for enterprise-level arrangements; I need something affordable for a one-person operator who consistently exceeds the current max plan + bonus credits.
@peakquant muted
Reason: 9. No spamming.
Duration: 4 minutes and 53 seconds
Closest thing I've seen to a mid-tier is the "Enterprise Pro" plan which is something like $40/month (more than Personal Pro, but less than Personal Max) - #💬│general message
Ask that to Any Ai as Well
It might be able to help you or suggest alternatives
I am wondering, do perplexity company make profits from their platform?
All That Information is Confidential
Any any Case The Valuation is 21B now
As long it can keep getting Funding it can be Sustained
This answer makes me worry more 😟
Because of The Recent Changes?
Because of my data is being recorded on western servers, america/idf can locate and assassinate me
I am important person in my environment 😢
Shouldn't give Confidential Data
Use Enterprise Plan if You Don't Want Your Data to be Given
That's great solution, thanks
Or Use Chinese Ai but then they get Your Data
Last Alternative is Local Deployment which is Very Technical
No escape
Sure I will
Make sure to Deep Research Thoroughly since is High Stakes
Not even a bit of your data should be stored or shared if is Really Confidential
What the hell
Qwen3.6 already?
bro u still write first letter of all words as capital
Why is no one fixing support issues for Enterprise MAX users? @tranquil orchid
thats crazy
its been like two weeks
curious what anyone here uses qwen for. I went through the whole ollama qwen and claude code install process and can't think of things to use it for. I'm also a vibecode noob
You use it locally?
yea
Which Model?
You could use AI to brainstorm, ask that model (or a cloud-based one) about yourself and have it come up with ideas that might interest you.
i bought code "12 months for free" and i wrote it on the web (stripe payment but without putting card and another info). I refresh page and i cant put my card with code because its "already used" wtf? how its used if i dont have pro plan on my account?
You bought from Third Party?
yes and it works at the first time but i refreshed web and preplexity ai support told me that i need to wait because system need to reload that canceled transaction
I mean it can fully replace claude code. Claude code is more than a vibe coder
It can do installations for you
I use it to setup comfyui, and other semi annoying installs
It grabs all the models, makes the workflows, etc...
You canceled the transaction?
no i cant because where?
Did you try to Reclaim it?
oh okay you are a bot
You bought a promotional code that was not meant for you, off an unauthorized 3rd party, therefore violating the ToS. The only channel to purchase a Perplexity subscription is through the website or app.
https://www.perplexity.ai/hub/legal/terms-of-service
You agree that Promotional Codes: (a) must be used in a lawful manner; (b) must be used for the intended audience and purpose; (c) may not be duplicated, sold or transferred in any manner, or made available by you to the general public (whether posted to a public forum, coupon collecting service, or otherwise), unless expressly permitted by the Company; (d) may be disabled or have additional conditions applied to them by the Company at any time for any reason without liability to the Company; (e) may only be used pursuant to the specific terms that the Company establishes for such Promotional Code; (f) are not valid for cash or other credits or points; and (g) may expire prior to your use.
Does Perplexity help generate reports, for instance, I work in construction management. will it format weekly site reports, etc.
depends on the complexity of the report, basic bar graphs can be done with regular chat but an analysis spreadsheet or PDF file you'd probably need at least Deep Research or Perplexity Computer.
𝐃𝐞𝐞𝐩 𝐃𝐢𝐯𝐞 ➛ [𝐡𝐞𝐫𝐞](#💬│general message)
안녕하세요, 저는 한국에서 온 학교 선생님입니다. 저희 학교 컴퓨터 실습실에 이 제품을 구매할 수 있는지 문의드리고 싶습니다.
please contact support@perplexity.ai
감사해요
Guys, do you use something to work with Perplexity in VS code? Im using the github connector but it just seems that lots of tokens are just wasted for perplexity to actually log to github and look through the files.
So idk, are there any ways to work in VSCode? Only PerplexiCode?
Is Perplexity positioned to support software development in teh same way that Claude is?
Is anyone else having issues accessing Perplexity.ai in the latest version of OpenAI’s Atlas browser? I can’t even get the login screen to load.
This is on a Mac M3 running the latest OS and updates, and Atlas is fully up to date. The Perplexity app works fine on the same machine, and I can access it through Safari without any issues. I’m also able to log in on my iPhone.
I’ve already tried incognito mode, clearing all site data, and restarting, but nothing has worked so far.
I was wanting to improve the extension so is there anything that i could add to it ?
I'm at the time where i need to study for my final exams so i can't update patchs and fixes for it
PerplexiCode is yours?
ye
Oh man im honoured ><
i have JUST learned about it, it seemed to be not as popular so i just thought im missing something out. It doesnt use API keys, just needing the pro acc?
yup
all u need
is just log ur pro account it
the agentic is a little bit behind
would update it better after i'm done with the tests
I mean you can'
can't
actually vibe code like a coding agent
but it can replica of 60%
where u can ask it about anything in ur database
the extension will find it and paste it to perplexity
along with the prompt + sys prompt
what i meant paste here is it communicate directly with pplx api endpoint
so
After u got the reply from perplexity
the extension will have option for u to either fix that file ( like when you vibe code )
or not
basically u just skip the middle point of using browser and manually uploading everything
Is it better / more productive than github connector feature?
Kinda killing me. Free qwen code extension does a much better job, so i figured lots of stuff is getting wasted in the process of goin to github/reading files etc
i haven't tested, it just basically a mvp app rn
as it took me
quite awhile
to find pplx api endpoint
and to actually use ur account in it
I will update it once i'm done with the test
Such data sources available only on Enterprise Max or regular Max too?
regular Max too
And not available in Pro?
Pro Perks are available exclusively to Perplexity Pro subscribers with a US-based subscription and are non-transferable.
Pro Perks are available exclusively to Perplexity Pro subscribers with a US-based subscription and are non-transferable.
Avoid excessive messages, all-caps shouting, mass pings, or repeating similar posts. Use channels appropriately.
and?
lmao
anyone else sign up for the pitch competition just for the free computer credits?
Give this a thumbs up if you've selected your stock for the pitch competition!
No! I am in it to win it
fair
Ye
Unavailable for non-US
I never said it was available elsewhere? I'm able to use them fine on a non U.S server

Me being 12 hours away from us
vpns are cool
Well they work on non U.S servers
nice
good luck
note: i haven't sleep for 2 days
uhh

doesn't seem too healthy
living on monster energy drink
idk but i start to constantly screaming like 5-10 minutes per time
...
idk
But anyways
concerning

I love uni life
where u can just
pull everything
in one night
( not sure if u will alive after that )
hmm that's interesting, I've seen some people say it's the opposite
depends on how you manage your time ig
it's more like a joke berd
not what i actually meant
It means that
U can either choose to be chill and wait for ur death
or
pull everything learned in 1 sem
into 1 night
...
avg uni life

i'm lost on the count of the energy drinks i drank
anyways
imma go and have some scream and continue studying for a bit
Our new model is now live on OpenRouter for an early preview, go give it a try! Looking forward to your feedback~😎
Quoting OpenRouter (@OpenRouter)
︀
Qwen 3.6 Plus Preview from @Alibaba_Qwen is live now for free for a limited time on OpenRouter!
︀︀
︀︀During this free period, prompts and completions will be collected and may be used to improve the model.
Im lost with their naming lol
There's a plus, an oss, a flash, and they replaced vl with omni now.
Is discord down
Or just me.
I haven't gotten the credits nor the prompt yet. Has anyone else gotten them?
I'm in the pitch competition

-# white monster?
can anyone help?
@young hollow simple yet effective
Damn
...
Has Perplexity de-valued their tokens? With exactly same morning usage, I am depleting the browser sidebar use 50% faster. Same prompt, same time, no interaction with Perplexity before. Don't have the "Computer". A Pro subscription is now less of a value vs previously.
Sounds like they are trying to push you towards the Max.
please I would like a chance to do this
What about pro for comet browser automation?
🗣️

scam
Any fellow maxxers setup a mobile battle station on their Perplexity Computer? Sorties in flight #runtheop
Interesting and Appreciated
@distant aurora they added another tool where gpt can put its suggestions for drafts in a box with quick copy paste and share buttons
damn, I love perplexity computer so much
I don't want to use any weaker ai model for complex tasks anymore, this is the baseline now
literally peak stuff, may it become even stronger
cute
Hello team my perplexity pro is not working in my phone and app not working
Can anyone help me claim perplexity ai pro
try with turning Web mode off
[Limits by Plan](#💬│general message)
please contact support@perplexity.ai
Hi folks! Is there any way to force Computer to use a particular (or at least the most powerful) model?
Hi
Same Issue here, is there a way to see when my limits going to be reset?
@dry kestrel
Let's goooo
what should i test
been quite awhile
returning to grok
you seriously randomly paid $300 for grok? lol
.... kinda ?
[Limits by Plan](#💬│general message)
Guys, how do I even use labs? Every time I try using it it switches me to Deep Research, thats quite frustrating
try those frontier math questions on it
curious how 16 dumbasses agents work on it
hi, I have Pro account and limited Deep Researches for almost 3 months now, I get like 2-3 per week and it's gone, saying I don't have enough credits to continue using Deep research, anyone had this problem?
I really like Perplexity but I feel like I get more with Free than Pro now, can anyone send a link where this discussion has already been solved, thanks!
keeps saying my monthly limit is reached, but it's been 3 months now saying that
Absolutely
U got rate limited
Just wait for it
and stop using deep research for a week
it should reset
Claude code leaked on x
I think you had in past as well ?
ye
it was
trash
Now i want to try it
again
no clue
Let us know The Test Results
💀

When did you become so Rich
It has a limit now and uses are regenerated - #💬│general message
No but i have gemini ultra and chatgpt pro
Claude code source code has been leaked via a map file in their npm registry!
Code: https://t.co/jBiMoOzt8G
Well uh…
That happened.
Its been quite few weeks now i dont see grok in perplexity pro do they removed it from there and yeah i havent seen not even once since it has dissapeared so its not about tokens or server busy thing
How do I get Perplexity Computer in my Mac desktop app? I can see it via web access.
I can't rewrite or switch models I am in pro plan, its showing me to upgrade to max,
Maybe you ran out of premium requests
Did you send over 200 in the past 24 hours?
What is happening with the pro plans, is there a limit now?
Yeah 200 in a 24 hour rolling window
Yeah.
No man I cannot use the rewrite button.
It showing that i have to upgrade to max.
Ah you might just be cooked then
@tranquil orchid Hey, I understand you have get as many max customers in this financial year, but try not to mess with pro users.
Thanks stella.
The issue with that is you're gonna nuke your context window
Specially if you were using Claude which has huge context windows compared to other models
I only use claude sonnet, other models are not good.
Real
GPT only knows how to hallucinate and write bad essays and Gemini takes 10 minutes to think of a mid response

How come perplexity is telling me that I can’t upload any more files? I have a pro account and I haven’t been using research or anything like that, all I’ve been doing is using the AI to help me with math, Aka uploading images of the given exercises?
Exactly, Sonnet 4.6 is GOAT.
They limited the usage for pro users
They reduced all usage when they released deep research and now pro usage limits are comically low
How much are we talking?
Like 5 images or something iirc
Cuz I used to use it way more like 4 months ago with basically no limit
Sonnet 5 soon! 🤯
I must have upload like max 70 images since Saturday
So I’m kinda shocked cuz it’s 20€ a month
Never had a problem with that before
Hopefully
have you even used those models? lol
Its better to take claude subscription now.
Sonnet 4.8 to be Precise
I try to test them all when the responses I get from Claude aren't good
Ironically I've been using nemotron for search
Also Opus 4.7 and Mythos 5
It's actually kinda good
gpt you should use 5.4 and gemini you should use 3.1 pro
I know there are no other options
also gemini reasons the least of the frontier models
it reasons so little that it can skip details...
Wasnt Gemini supposed to be like really good at deep context reasoning
it can hold data in context but actually working well is a very different story
Well I guess that’s it for me in regards to using perplexity
Gemini have a lot of capabilities but it is stupid, Chatgpt is kind of stoned, they are good at coding but not general research and daily use. Claude is best at everything.
I haven't tried Codex
Because I already pay $200 for Claude 20x and I don't wanna pay another $200 to try codex
you can do a heck of alot of work with $20 codex
That's Codex mini tho
the rate limits are much higher than claude code
Full codex is exclusive to the $200 plan
no you can use all models but lower usage rates
Codex is good but claude code is way better.
Yeah I use Claude code for everything
Hmmm from what I've read the $20 only gives you access to Codex mini
user intent and context understanding is something every other model lacks.
Try Free Tier
You might be amazed
I have seen many use it and be satisfied
Hmmm
Free Codex Limits are Far Better then ChatGPT
Is GSD on codex?
Yes
Claude code leaked mirrored to Gitlawb. will never be taken down in decentralized git platform. repo link in reply
See is only 2x
In comparison to Free/Go
Free Tier is Viable for Now
i mean its still more than claude lol
compared to their $20 sub
Precisely
codex is compute hungry so cant blame them at all
I have seen Claude Max Users Burning Their Quotas Fast
this is so not true
Every interaction with AI is getting trained except localmodals, even if they are saying it won't be used for training, they do.
Many of my Friends Use Free Codex for Coding
There are ZDR policies that they have to comply with if you use the API or are enterprise user.
They are not power users tho but they use it well
It's extremely limited though
That's why i said Viable
Also this is limited time offer
you can't do much on the free tier
I don't know about your experience but mine and my friend's has been satisfactory
Of course we are not power users
yeah well I've burned close to a billion tokens for this past month
That Makes Sense
just codex on pro?
damn
They have certain clauses in their policies for the work around, they might say something like in order for model safety they will be moderating the model response, when that moderation happens our data gets out.
was making an extension + a bunch of other side projects
Daym that's Cool
@bhavishyaaaaaa_09955 muted
Reason: 9. No spamming.
Duration: 4 minutes and 53 seconds
Does anyone know why when i give a prompt, the response stops halfway through?
In need of team member response- I’ve been told someone would get back to me in 2 business days for the past 2 weeks
Has anyone seen anything in writing about Personal Computer? I have only seen the video.
newbie here and a small healthcare marketing and sales consult...using ai for marketing and sales- sales, marketing, messaging, research and tools for productivity...what am i missing specially in a business model that is supposed to be a human rights venture
Yes. I do that, and then it says “it can be done with Deep Research so im switching on it”
Anthropic is now officially more open than OpenAI.
@alluring_mango_41438 muted
Reason: 9. No spamming.
Duration: 4 minutes and 53 seconds
Has anyone else evaluated responses when switching between models in the midst of a thread? For example, generating a response with gemini and then switching models to claude in the same thread and asking for self-reflection. (Sort of a poor man's model council). Is this worth the effort?
https://www.perplexity.ai/hub/blog/everything-is-computer and a short blurb where you can join the waitlist here: https://www.perplexity.ai/personal-computer-waitlist
There are other third-party articles on it, but those two are on perplexity's main site at least.
anyone have a negative balance in computer all of a sudden?
please report here: https://discord.com/channels/1047197230748151888/1140622008086970420
how to format your report: #1358926750108614928 message
thanks!
sounds like an undocumented feature 🙂 though I would hope they don't ask for back payment
reported, last balance was positive and all of a sudden, double negative balance....
fetch('/rest/billing/credits/balance?...')
.then(r => r.json())
.then(data => console.log({
balance_cents: data.balance_cents
}));
// Result: balance_cents is negative even when visible usage is zero
that provides a detailed description of usage,balance and you can query per tasks credit consumption,
But seemingly the backend a calculation issue
Not always the case
Especially not with grok lol
@dry kestrel think mythos is anthropic their pro tier model?
Like 5.4 pro and deep think?
They were the only one without an ultra heavy model 
They would even release anthropic source not to publish the list
What a world we live in
perhaps
it seems more like GPT-4.5 though
It will be even more expensive then the ones you mentioned
My attempt at prompt engineering didnt work 
I run out of 45k computer tokens in 3 days
gonna grab some strangers iphone on the street
Giant pre trained model?
Did anyone get into their Perplexity startup program?
hi
@notoree muted
Reason: 9. No spamming.
Duration: 4 minutes and 53 seconds
Eh, we don't know the pricing for Gemini 3.1 Deep Think yet
Is it possible to see which model answered your query when you have "best" selected?
@ah_mady muted
Reason: 9. No spamming.
Duration: 4 minutes and 54 seconds
No
wouldnt be surprised if sonar happens to be the "best" model for 99% of my queries then lol
@ajajay799 muted
Reason: 9. No spamming.
Duration: 4 minutes and 54 seconds
it's usually GPT-5.1
at least for the simpler queries
@fralt007 muted
Reason: 9. No spamming.
Duration: 4 minutes and 53 seconds
@green crystal bro keeps a grudge journal now. https://cunnyx.com/i/status/2038995503141065145
Codex with GPT 5.4 High has been absolutely crushing it today.
︀︀
︀︀Spawning GPT 5.4 sub-agents across X, YouTube, Reddit, TikTok research tasks.
︀︀
︀︀All in parallel. No rate limits. No 529 errors.
︀︀
︀︀This is what a reliable coding agent feels like.
︀︀
︀︀But OpenAI is cutting Codex rate limits in half in 2 days.
︀︀
︀︀I just left Claude Code because Anthropic slashed limits without warning.
︀︀
︀︀If OpenAI does the same thing, where do we even go?
︀︀
︀︀Every AI company gives you the product.
︀︀
︀︀Then takes it away.
Hellow everyone
Hii
Bruh the server is dead 🥲🤧
Idk 😶
I used to be here and enjoy, the chat didn't use to stop and now , this much silence
Let's join LMarena
@dry kestrel
Do u know
grok
context windows ?
I'm getting annoyed when i have to manually type next for every chapter and then copy it\
Heyy
hi
Hi
How do I get a refund for perplexity?
Will the unsupported region for image generation ever be fixed?
Startup program?
Buy max
It's fixed on max
Peak
This is getting interesting
Ye i did
Got my enterprise pro
Idk grok 4.2 heavy seems great
But the context windows
Bro
I'm buying second perplexity maz for tokens
😹
How did I run out of 45k so fast lol
Hi guys, just wanted to confirm with u guys if perplexity computer feature is available on mobile app right? I am facing issue that suddenly I couldn’t use perplexity computer on mobile app and reached out to support team. They said that the feature is only available on Desktop
mg
Hi I have joined the server to ask if anyone knows if the features for pro have been downgraded. In the last couple of days I've been told I've hit a daily file upload cap which has never happened before I have in the past uploaded way more documents. And just now I tried to use research mode and it said I had to upgrade but I am already a pro user. I have never had problems with Perplexity before so I was wondering if anyone knows if they are downgrading the capabilties or pro to push people to buy max. I cannot afford to spend max money.
Hi there! My Perplexity comet stoped working. I also can not reach their web site through chrome. Does anybody knows where is the problem and how can I fix it?
Limits ➛ [here](#💬│general message)
Enterprise
is a solid upgrade, you don’t need 
thanks loads for replying, I saw that post was dated August 2025. I cannot understand why I am only hitting blocks to my useage now as I've not increased my useage particually at all in the past few days and these issues as I say have been happening for the past few days which was why I was wondering about if the caps are new or if there were set with the latest update. Seems to have started after the new deep research feature was launched.
recently introduced (new limits), it coincided with SOTA Deep Research launch
https://www.perplexity.ai/page/perplexity-upgrades-deep-resea-bHxtGLcaQeCpJSmq.DYN8g
Ah thanks again, yes they aren't going to let me use Opus 4.5 with my £20 a month pro fee are they. I understand.
I'll go and shop around, it maybe that pro is still the best in class for the price point not sure yet. But no harm in looking. I got pro to use the research mode you see that's why I'm taking it so hard. Shame.
@wonderking_op muted
Reason: 9. No spamming.
Duration: 4 minutes and 54 seconds
Does anybody reached out to perplexity web sites?
Sonnet Thinking is way to go, in your situation
thank you this is brilliant!
𝐃𝐞𝐞𝐩 𝐃𝐢𝐯𝐞 ➛ [𝐡𝐞𝐫𝐞](#💬│general message)
🤿
Does anyone know the Reason Effort for All Models of The Pro Plan?
uhhh no clue
From last i remember I think it's Medium
I don't know about Latest One
i kinda
love grok now
grok
4.2
is so much better
than their previous 4
Daym
Finally is Good
How many tests did you do?
Rate it out of 10
For the response speed ?
Ye totally
10/10
Love it
for the quality
i'd say 9/10
for the context windows
6/10
Claude and gemini or chatgpt
can do what i asked for
here i have to manually type next for every chapter
and
no pdf file
Ye
So what's Your Ranking of Models and Platforms?
Ufff
Doesn't GPT have Unlimited Limits
More Upload is Valuable
The context windows is what u consider
and yet
the reponse time
i ain't waiting 30m
Ohhh so GPT is Slower
just to learn basic matrix
Crap
Not
the case
Grok heavy run on multi agent
to be specific
16 agents
So the reply time is 16x faster
16x Faster






