💬│general

[twin grail]

My claude just got x4 limit

[gloomy cipher]

i might go back to kimi code sub but i still dunno if its worth

over the deepseek

[twin grail]

It’s

Deepseek is way too hallucinating

It’s not reliable

For heavy coding and long sessions one

Kimi did a great job for this

[gloomy cipher]

its doing very well for me but that might be because of the harness iam using

it has like idk 12 special agents or something like that

[twin grail]

Does deepseek filter heavily on security stuffs ?

[serene sand]

@waxen wing

https://www.perplexity.ai/help-center/en/articles/13901296-how-credits-work-for-enterprise-organizations-admin-guide.html

https://www.perplexity.ai/help-center/en/articles/13901321-how-credits-work-for-enterprise-organizations-member-guide.html

https://www.perplexity.ai/help-center/en/collections/18789398-computer-for-enterprise.html

[twin grail]

Because i was able to use Kimi for my security research

It worked perfectly

That’s why i’m loving it atm

[waxen wing]

Is it possible to talk to a real person at Perplexity?

[twin grail]

From what i experienced

When i used to have enterprise subscription

U can talk to them

But it heavily depends on ur problem

Like if the problem is one of the main thing ( like bugs that people reporting constantly and requiring a solution)

It might take them awhile to reply to u

Even for me

[fallow notch]

Yoo guys what’s chill

[twin grail]

Used to be an enterprise max user

Took them a month to reply back about the campus program

So ye

Billing should take less than newly problem

Like computers as they will need to ask confirmation from the higher up

Which could take around 14 days ( average from my experience)

[frosty patio]

Did you buy a code (such as from g2a)?

[green crystal]

@dry kestrel https://x.com/i/status/2054958163925852416

leo 🐾 (@synthwavedd)

For everyone asking about GPT-5.6's frontend abilities - it still sucks. Nothing is certain in life, except death, taxes, and GPT models generating the sloppiest UIs of all time.

Gemini 3.2 Pro is heading in the same direction too, regressing versus 3.1 Pro.

X · May 14, 2026, 4:13 PM

This can't be real right
How they using 5.6

[dim radish]

U meant me 🤭 well, I'm a girl from thailand btw yaaaaa

[frosty patio]

Sawatikhaaa

[dim radish]

Aww are you Korean? Annyeongggg

[frosty patio]

wkwkwk no the pic is my gf from indonesia (Chindo)

Just likes idol makeup

[dim radish]

Then hai apa kabarmu? 😜😅

I know Indonesian too jugakkkk just in case wkwkwkwkwkkwkw

[frosty patio]

I don't, even my chinese is better

[wooden parrot]

Hello everyone

[frosty patio]

Hi how are you

[honest horizon]

Is this against pplx TOS?

[frosty patio]

I think so, thus they are requesting phone verification from the original purchase country

[honest horizon]

yeah true, though I don’t think that stops global accounts being sold

[frosty patio]

atp I think the solution might be like
People complain/ request refund to g2a --> g2a forces sellers to not mark global anymore and limit to the country (and maybe refund)

[honest horizon]

that’ll probably be the case

[agile wharf]

Enterprise Max Payment Issue – Need Help

Has anyone experienced this issue with Perplexity Enterprise Max?

Problem:

Subscribed to Enterprise Max plan (~$400/month)
Stripe shows payment successful (invoices received and paid)
Dashboard still shows red banner: "Complete payment procedure"
Team blocked from accessing Enterprise features
Timeline:

Payment confirmed: March 30 + April 11
Support ticket opened: May 13 (24h response promised)
Still waiting for resolution
Question: Is this a known sync bug between Stripe and Perplexity backend? Any similar experiences or suggestions?

[frosty patio]

I will keep suggesting revolut, it's $8 less and gets you a nice card + perplexity pro + nordvpn

[honest horizon]

It’s a shame Perplexity is cutting down on Pro usage. They used to give out Pro subs basically to anyone but that backfired so now they force feed you Computer

that’s what I think is happening

lucky that I got a good deal because I’m not sure if I’d be paying monthly for the Pro sub right now

[serene sand]

@signal hamlet

[wooden parrot]

I’m good and you

[frosty patio]

Yeah as I wrote about above it's annoying to have what is essentially an advertisement suggested as a response, with things like "make a pdf" tacked on for no reason, seemingly to justify computer usage... Even though, gpt and gemeni can run scripts and make pdfs too

[frosty patio]

better than I deserve!

[honest horizon]

yeah it’s quite annoying. And almost all of the time it’s 100% Computer suggestions now

[agile wharf]

Update: Support replied via automated bot promising resolution within 24h. It's been 3 days now - no action taken, team still blocked. The bot response was completely generic with zero actual troubleshooting.

[wooden parrot]

I’m happy for you cause I didn’t get what I deserve

[honest horizon]

he’s your best bet other than support ^

[frosty patio]

I do see where perplexity is coming from though, like all AI "unlimited" subscriptions it's probably a 300%+ loss on average and perplexity sold it for much less through deals like revolut and wherever the g2a keys are coming from. And most AI services copied perplexities scraping tech which was their original edge over e.g. chatGPT. Running AI is unprofitable and very reliant on huge investor bucks.

[frosty patio]

What's wrong?

[honest horizon]

absolutely right, that’s why Claude and ChatGPT have upped their prices too. It’s a shame for people with yearly subs though. You pay for something upfront and then it slowly gets taken away from you throughout the year

though I personally can’t complain as I basically got it for free a few months ago. But I know how frustrating this is for others

[frosty patio]

Yeah but the real big shame is that the bubble might be shifted on to the general public before it pops. Stock indexes like NASDAQ used to require that a company be public for one year before being included in the official index. They are changing this so that openAI and anthropic can be included very quickly (month or less)

This means that when they IPO, index funds, which almost everyone worldwide that have money invested owns, which makes up 50%+ of ALL stock ownership will be forced to buy a small number of openAI and anthropic stocks, at whatever market price is.

This will include national funds, pension funds and such that follow the indexes

[halcyon steppe]

Hello there, I am on the pro plan, however I have run out of the deep research credits for the month, is there any possible way that I can top those credits up, I don't really have the money to get max

[serene sand]

Limits ➛ [here](#💬│general message)

check Enterprise

[honest horizon]

shame pplx has stayed quiet about all this

[frosty patio]

Index funds are already MASSIVELY investested in Nvidia that will tank if AI companies are exposed as unprofitable. With openAI and anthropic it will be massive. The general public, especially unaware pensioners, average beginners, people that never agreed at all might be the ones to pay for all the AI in the end

[honest horizon]

I think you can see what users are thinking based off of the announcement post reaction lol

[honest horizon]

unfortunately most things like these mainly impact the “poor”

[frosty patio]

Yeah. Relatively poor, anyway. I haven't figured out the best hedge against this obvious vulnerability yet. Bet i have is bear NVDA x2, but can't really short openAI and anthropic until they are listed...

I hope perplexity can get a nice big investment to keep pro good and remove computer adverts, to last us until it pops!

[honest horizon]

I don’t think it’ll be wise shorting right after launch

all “exciting” IPO’s tend to do well in the first 3-5 days

and after that the hype dies down and price goes either way

[frosty patio]

Indeed. and then bellow market for the rest of their lifetime

Short after index funds buy in (Looks like it will be 1 month)

[honest horizon]

took a nice short with Figma just like that

[honest horizon]

this week seems like a good week for shorting ai

[frosty patio]

topkeke

[honest horizon]

I’ve sold most of my Intel this week too

[frosty patio]

But im not planning to do some yolo wallstreetbets, i just want to hedge, so for example if I have $1000 in a global index fund, NVDA is 5.5% of that (Isn't that crazy?!) so $55 ($22.5 with 2x leverage) in NVDA shorts

If im totally wrong it's okay, at most a 2.75% loss

[fresh yacht]

Can we PLEASE have a separate topic channel "CREDITS'? This clearly is a huge issue amongst all of us but as a newbie this discord is incredibly useless in regards to finding information if every second message reads 'I ran out of credits, please help' or 'I bought credits and their not showing up' etc!!!!

[halcyon steppe]

I have a question, if I buy Computer credits, can those be used so I can continue using deep research mode?

[heavy portal]

since credits only apply to Computer (so far), the next best thing might be the https://discord.com/channels/1047197230748151888/1483514338345091213 channel.

[glass sparrow]

Does anyone here have experience with AI tokens? I'm interested in learning more about their usage and value.

[tranquil orchid]

@yxshhhh69 muted

Reason: 9. No spamming.
Duration: 4 minutes and 48 seconds

[sharp granite]

perplexity pretty ass if you dont pay

[sleek tiger]

Hey Grok, Grok Build just came out today in early beta. Is it possible to connect it to Hermes

[sharp granite]

hey grok

[sacred current]

yo

[distant aurora]

@dry kestrel https://cunnyx.com/i/status/2054918374837322140

Kimi.ai (@Kimi_Moonshot)

Meet Kimi Web Bridge - Kimi's browser extension.
︀︀
︀︀Agent can now interact with websites like a human: search, scroll, click, type and complete tasks.
︀︀
︀︀Supports Kimi Code CLI, Claude Code, Cursor, Codex, Hermes, and more.
︀︀
︀︀Available now on kimi.com/features/webbridge and the Chrome Web Store.

**💬 80 🔁 162 ❤️ 1.5K 👁️ 126.0K **

▶ Play video

CunnyX · May 14, 2026, 1:34 PM

[dry kestrel]

yep been a thing for a while https://x.com/eddiboi/status/2046598585806950673

EdDiboi (@eddiboi)

Kimi WebBridge spotted! Grab the browser extension that links your Kimi desktop app and lets the AI agent control the open web locally for automation.

https://t.co/WOuzVPGqO4

▶ Play video

X · Apr 21, 2026, 2:35 PM

very cool little extension

and desktop app

[heavy portal]

This makes me want PPLX to fully switch to an only-credits system (with monthly allowances) even more.

[sharp widget]

⬆️

[twilit crown]

We are cooked

[heavy portal]

The guy from Chicago not likely to be gone any time soon.

[echo stone]

a better headline might be certain AI chatbot users can get their delusions reinforced.

[cunning pebble]

I keep getting a message directing me to a subscription screen. I can get out of it and it works fine but not sure why it's doing that as I'm on a pro plan.

[heavy portal]

is it asking you to subscribe, or to subscribe to Max specifically?

[cunning pebble]

I forget now. It doesn't come every time, maybe just once every day or two. It's a bit annoying. It's only happened a couple of times, never had it until recently.

[heavy portal]

You might be hitting the limit, then later when you come back it has regenerated more usages - #💬│general message

If it's Pro Searches (i.e. picking a specific model) it might be switching back to the "Best" mode model when you come back to it after closing that subscribe screen. "Best" mode prompts are unlimited.

[lapis hamlet]

these assholes keep adding great stuff for computers but most subscribers are pro adn pro is not able to use computer credict(not given)

[normal mirage]

@fresh yacht

@normal mirage

[minor shard]

Niceee!

[distant aurora]

Does it run on the limits of the sub or have its own pricing system?

https://cunnyx.com/i/status/2055083187429355823

Chetaslua (@chetaslua)

🚨 Polymarket also knew today that it would be Gemini 3.5
︀︀
︀︀30 hours ago it was at 3%
︀︀
︀︀New checkpoints are crazy good , we are not allowed to share output yet , I am trying my best to get something for you guys

Quoting leo 🐾 (@synthwavedd)
︀
3.2 out, 3.5 in
︀︀
︀︀interesting

**💬 14 🔁 7 ❤️ 232 👁️ 23.1K **

CunnyX · May 15, 2026, 12:29 AM

https://cunnyx.com/i/status/2055109194043494654

Chetaslua (@chetaslua)

🚨 Gemini 3.5 output
︀︀
︀︀Join Discord server link is comments
︀︀> Most of the news will be there in no time
︀︀> Good community
︀︀> Most important sometimes they give access to the server member

Quoting Lentils (@Lentils80)
︀
Gemini 3.2 got renamed to 3.5 👀
︀︀
︀︀In other news, below are outputs from a strong Gemini 3.5 Pro checkpoint called "Cappuccino"

**💬 8 🔁 4 ❤️ 71 👁️ 5.7K **

▶ Play video

CunnyX · May 15, 2026, 2:13 AM

[frank kelp]

so i just discovered that once my perplexity computer chat gets long, it starts to hang - the suggested solution by the app is to generate a 'context' file to paste into a fresh new chat. This is quite frustrating when it could just go back and wipe any unnecessary history data from the current chat and we continue on without me having to start a whole new conversation. Is this something that is being worked on?

[dry kestrel]

limits

[lapis hamlet]

it's 12:38 am

i am so fucked

[shrewd cradle]

what exactly do Pro users even have? I"ve been so confused the past few days and only now ran out for my deep search. So even the specific models run out too? Do we have a credit system to keep track or is it just gonna be disabled if I go over it

[rugged trench]

No more 1 year of Per for free?

[distant aurora]

https://cunnyx.com/i/status/2055082821979443426

Just a dragon (@Waguri_Kaoruko8)

The new Gemini Spark model will have Agent mode / Chat mode.
︀︀
︀︀New advanced use of tools.

**💬 2 🔁 1 ❤️ 26 👁️ 3.9K **

CunnyX · May 15, 2026, 12:28 AM

[main wind]

shhhh passing the 8 minute mark 🙂

ah it didnt eror out but it stopped just the timer was running

[distant aurora]

@green crystal nothing beats gpt thinking search and grok multi agents search.

Closely followed by Deepseek and maybe Gemini.

[mighty dove]

Ghuyz

Is anyone here?

This is MW

[main wind]

im here 😉

[glad jacinth]

Hello everyone, Pro subscriber here. I've been using Perplexity since December and I never had issues with it. I use it for storywriting purposes. Since a couple of days, I'm no longer able to use any other model than the Best one, which doesn't generate the kind of output I'd like (I tend to prefer Claude Sonnet, because it has better consistency at following the previous prompts to keep the stories going). On the phone app, it just reverts to the Best model while on my PC, it says I should upgrade to Max.

From what I've seen I believe I might have reached the limit and I need for it to reset, but so far I never had this issue. Does anybody know how much time does Perplexity take in order to reset the usage limit? So I can have a better idea about how much I should wait. In the past days I had this issue only when I regenerated a prompt, but if I edited an existing one, it worked fine. Now I'm no longer able to do that, neither on phone or on PC.

[serene sand]

Limits ➛ [here](#💬│general message)

Regeneration rate ➛ [here](#💬│general message)

To reduce repetitive uploads and improve organization, store frequently used files and images in Spaces

https://www.perplexity.ai/help-center/en/articles/10354807-file-uploads.html

[unborn crane]

Hi

[midnight narwhal]

Hello,
I’m trying to get in touch with someone from the Perplexity marketing team. I’ve sent a few requests, but so far I’ve only received automated email responses from Ai agent saying my inquiries have been logged.

[unborn crane]

How to get pro role if I’m pro user?

[crystal dune]

hey

[serene sand]

@unborn crane

[neat warren]

can anyone help me - https://discord.com/channels/1047197230748151888/1504459873801207880

[serene sand]

Please send all the details to our [e-mail] team, they will check and help you with this.

Be patient and wait for a response, the team needs time to review and then respond to tickets, and there are a lot of tickets in the backlog right now.

[midnight narwhal]

"Hi wlyderaf,

I hope you're doing well. I am looking to connect with someone regarding a B2B cooperation between Perplexity and my company.

We would like to get information on how to purchase licenses and how to administer them for our users. Could you please provide a relevant email contact where we can discuss these details and ask a few questions?

Thank you for your help!

[green crystal]

yeah both can seriously dig into info

[keen kite]

Business lesson, don’t continue reaching out and sign up for another provider as Anthropic or OpenAI instead

[crystal dune]

uhmm, i’m doing great. hope you’re good too?

[serene sand]

please contact [e-mail]

[midnight narwhal]

We have already sent three emails to [e-mail], but we have received no response for over 10 days . This lack of basic communication is unacceptable for us. I have reached the limit of my patience and will now be looking for alternative enterprise solutions. Good luck with such great bussines model.

[serene sand]

Most of the info you need is located here: https://www.perplexity.ai/help-center/en/collections/11312901-perplexity-enterprise.html

[lusty harbor]

Perplexity seems to have stopped working on my iPad mini 6.
It starts up, it looks fine, it doesn’t respond to any touches on the screen.
It’s just, stuck..

I have tried offloading the app, and reloading it. I have tried removing the app and reinstalling it.

[sleek loom]

Why does perplexity now require subscription to talk to it even if it’s just general inquiries, base model.

[wooden parrot]

Can I send you a dm?

[lusty harbor]

Perplexity on my iPhone is helping me to debug perplexity on my iPad. It’s a strange situation.
The last iPad app update is from just 14 hours ago. It was working fine for me last night, it must have updated in the background before this morning, now it doesn’t respond to any touches, even though the logo keeps animating.
I have done another remove app, full iPad shutdown and restart, reload app from App Store cycle, and it still won’t respond to any touches.

It’s looking less like ‘bug fixes and improvements’, and more like ‘bug introductions and total lockout of all functionality’. 😏

[clear canopy]

Hello! How's everyone doing?

[cursive thorn]

Does anybody know why perplexity on Singapore app store is 18+?

Have to vertify credit card, really frustrating for oversea student…now I can't install perplexity app anymore

[tiny cedar]

[worn bison]

Why is this happening after I carefully managed my pro query tokens?

[frosty patio]

I don't know but sg has so many rules and regulations so it wouldnt surprise me if thats why

[abstract sentinel]

does someone know how to fix this?

[oak meteor]

Get an Indian number

[fallow notch]

Let me have a look

[lone trellis]

I’m feed up of computer updates while pro user only had one new feature: extreme rate limits and the coupon breaker phone prompt

[twilit crown]

Why is Perplexity Computer so forced

Like Pro users get no credits but they beg you to use it

[green crystal]

its so you buy those credits lol

you're subscribing for the "honor" of purchasing credits

[twilit crown]

Who actually buys them though

[grand osprey]

whats that name 🥀

[grand osprey]

whats the coupon

[cyan steppe]

Hey guys. Who’s up for me referring them to Revolut (there’s €40 to be won)?

[tranquil orchid]

@ilham.jr muted

Reason: 9. No spamming.
Duration: 4 minutes and 51 seconds

@parthpp muted

Reason: 9. No spamming.
Duration: 4 minutes and 52 seconds

[quiet lagoon]

Seeing a lot of this on reddit

[distant aurora]

Hope

[errant tundra]

What is wrong with Deep's research again? Their effort level seems to be downgraded to 30%

I ask Deep Research to write an analysis, and it barely gets a third of the job done

[quiet lagoon]

Yeah used to be a lot better quality wise

[errant tundra]

Yeah, first it gets dumber, now it just don't do the work anymore

I send it nearly 3,000 words of prompt. It gives me back less than 1,500 words of analysis. That's ridiculous

[serene sand]

let’s check, drop your prompt/thread link here if you can share

[shrewd cradle]

what counts as a use of lab credit?

[heavy portal]

Nothing, they removed that feature and moved the functionality into Deep Research and Computer

[shrewd cradle]

so generating things would just require deep research? and pro queries?

[quiet frigate]

As a Pro use, why am I asked to upgrade to the Perplexity app on the Mac and use Perplexity Computer, but then it says I have no credits. What’s the point!

I got fed up and deleted the app. Don’t send me emails saying Computer is available to all and then say I have no credits to use it. If it’s not available to use without credits, then don’t mislead people into thinking it’s available.

[grand osprey]

holy tuff

https://tenor.com/view/love-i-love-you-love-me-you-dont-love-me-why-dont-you-love-me-gif-13050251675719002507

[cerulean garden]

Hello
I got a promo code from fireflies today
And i tried it on perplexity and its appear as expired

[slow wing]

AI is an idiot

AI AI AI AI ALL HAIL AI!!!!!!!

[stoic turtle]

#💬│general is everyone facing less pro searches issue? I haven't used pro searches continuously and still it is showing that I have only 3 advanced searches left.

[lethal marsh]

PRO User: region not supported for phone verification help me with this

[red geode]

This is bullshit. What’s the point of paying for PRO if I can’t access any model besides sonar 2.

[shrewd valve]

Same here, checking the API rate limits you gain about 1 pro search per 4 hours or so, pointless

[vale temple]

Just me or is iPad app unusable

[split island]

why does this happen https://discord.com/channels/1047197230748151888/1504261306696335592

https://discord.com/channels/1047197230748151888/1504632270374309990

it's bugging me

can somebody help?

[winter pond]

Perplexity has taken 2 weeks to get back to me on email this is outrageous, I want my refund for the max subscription because it doesn’t even work

There has to be some kind of offense for that because there’s no way they can do this and just get away with it

[winter pond]

What’s the point in paying for max when I can’t even use the one thing i bought it for

[forest rock]

i get any bonus for bug reporting?

[winter pond]

No and they could care less about what you report

[forest rock]

that sucks, i only report on application that gives bonus

is not acceptable no MCP integration with Windows

[calm void]

You're absolutely right! Any solution?

[forest rock]

Sonnet from perplexity more dumb than Haiku from anthropic

[red geode]

Pro used to be so good

[red geode]

L perplexity

[forest rock]

in my opinion perplexity is a firm with a high risk of bankrupt, literally there is nothing that makes it unique, even the models are not smart as their original provider
they got billions and turned this application useful to "Mac" only. Very very distant from real users

[red geode]

Turning to competitors

[red geode]

Truth nuke

[forest rock]

also Sonar is very distant from real money, SWE automation and others thing. that's why they are removing the subsidized based subscription.
hahaha

[wide pasture]

what the helly i thought it was pretty much infinite

[red geode]

Welcome to the club

[wide pasture]

is it gonna reseton monday

[fresh yacht]

I am having continued stalls within Perplexity where it says it's doing something but clearly it isn't. What's the best way to re-ignite it doing what I asked it to do?

[fallow badge]

I'm trying out superhuman and Fyxer.AI. Is there a way I can use Perplexity to handle my mail (auto label and write draft replies)? I'm using Perplexity for more and more these days and dont really like to use more apps and subscriptions.

[placid hatch]

Someone play to Vintage Story Civilizations ?

[honest horizon]

huh?? how so

[dim vapor]

I have the perplexity pro version, how do I get the API creds ?
I want to use claude model in my visual studio code?
Can anyone please guide me here

[honest horizon]

@serene sand

[heavy portal]

Click username in lower left, then All Settings at the top, then scroll down the left side to API Platform

Pro no longer receives $5/month API credits, but that's how you access it.

[cunning rock]

Hello perplexity team,

I am considering upgrading to Perplexity Max in the near future (depending on personal budget and costs required to do so); in terms of credits how much do you recieve per month on Perplexity Max and how much is credit usage per text generation? I tried to figure out the consumption of my credits for use of computer based on what is available on the usage and credits section but I don't have all the necessary data to make that judgement call currently

[heavy portal]

Here are some details - #💬│general message

[cunning rock]

It doesn't say the credits/month for the max plan; unless it is still the 500 computer credits/month? I thought it said somewhere that it was 10,000/month but I am not certain if that's the case

[hearty mango]

Hello 👋

[serene sand]

@cunning rock

[tranquil orchid]

@rachit_oo7 muted

Reason: 9. No spamming.
Duration: 4 minutes and 50 seconds

[cunning rock]

Cheers for that! I appreciate it! Do you know for how long it will be active for? I am just adjusting for some new changes and budgeting so can't comforatably adjust to max straight away

[serene sand]

one month

[wary sonnet]

I am a pro customer. What is pro discord and how to access it?

[heavy portal]

#💬│general message

[serene sand]

if you where asking about promotion, idk

bonus credits last 30 days @cunning rock

[red geode]

Idk maybe Sunday

[fallow notch]

Oh I see

[naive meteor]

Hey guys, can someone explain in more detail (some examples would be really appreciated as well) how the credits work, and if I paid for individual Pro for the entire year, how could I switch to individual Max, do I have to pay the full Max price or it can be adjusted based on how much I used out of the yearly subscription and I will pay the difference?

[dry kestrel]

It's prorated

[gray portal]

Hello ppls hope everyone's doing great. I had a doubt now, these days I am getting popups saying you have X amount of queries left for this month to use advanced AI, as it was not the case before, but why now, when I don't use it that often?

[crimson yacht]

Hey! I have doubt I have pro and says that I do not have more credits. But it's also saying that the next month I am gonna have also 0 credits. Does not make sense 🤔 so I am not gonna be able to use anymore perplexity?

[short narwhal]

Pro users don't get monthly credits. And you only need credits for Computer. You can still use search, research, and other features

[sly musk]

God damn

[wheat trellis]

whatever u guys did from 200 Pro messages limit, mine did not get replenished.
I thought this is replenished per nth seconds? [Example: I got 166 messages left but goes up when I don't use it] This change just only happened to me just now or I'm affected from your A/B testing.

definitely bruh moment, from 600 to 200 to literally hard cap 200 per week. This pricing is ridiculous gng

[civic wolf]

Something strange has happened to my perplexity iPad app - it has stopped accepting any touch screen input. I’ve removed and reinstalled the app, reset the iPad, and tried an Apple Pencil but no joy. The app comes up and sits there, and tapping in the input area doesn’t bring up a keyboard either. All other apps are working fine. Anyone ever heard of this? The same app works fine on my other iPad both on iOS 26.5. I’m just wondering if anyone has seen similar behavior before I fill out a bug report.

[mild cedar]

where did you get this dashboard from? 😮

[green crystal]

@distant aurora i wish oai didnt remove study mode from mobile and kept it more obvious in the selector menu
5.5 with study mode on is incredibly good lol

[distant aurora]

You can make a folder with study mode type instruction maybe.

[tawny wharf]

Perplexity Pro has a weaker memory system than when using models directly, like when using Claude projects. With the free tier as well.

Context and coherence are many times lost, even if using spaces with added source files and instructions within the space.

It’s simply NOT excellent in a professional use case.

Also I’ve found that generating pdf files create errors like text overflow, especially inside forms and tables. Generating the same pdf directly with Gemini, outputs a flawless document.

Hopefully it improves with time..

[distant aurora]

@green crystal @gloomy cipher @dry kestrel https://cunnyx.com/i/status/2055475929766629690

Chetaslua (@chetaslua)

🚨 Gemini 3.5 Flash on @arena → Minecraft in a single prompt.
︀︀
︀︀Best implementation of minecraft ever seen by any ai model till date for one shot
︀︀
︀︀What actually surprised me:
︀︀>zero external assets (no .png textures, no models)
︀︀>working block place + break with hit feedback
︀︀>procedural textures drawn in code, not loaded
︀︀
︀︀Arena link in reply to verify by yourself

Quoting Chetaslua (@chetaslua)
︀
🚨 Gemini 3.5 Flash checkpoint notes:
︀︀
︀︀The old “cool Gemini UI” vibe feels mostly gone. It is more GPTified now.
︀︀
︀︀Good: fast, smart for a Flash model.
︀︀
︀︀Bad: it keeps filling websites with HUDs/panels, even when the prompt asks for restraint.
…

▶ Play video

CunnyX · May 16, 2026, 2:30 AM

[dry kestrel]

yes very interesting

[green crystal]

so pretty good but bad instruction following
and double the strike if it also still hallucinates as much lol

[proud terrace]

I'm gonna leave the perplexity team preggers

And diddle away

Just because

.

Doesn't matter what gender they are either

They will be done preggers regardless

Nobody is standing on my way

[gloomy cipher]

Tbh the instruction is just 3 words

[rotund dune]

The perplexity app on my iPad stopped working yesterday. It opens, but i can’t write och click on anything, it freezes. I tryed to restart the iPad and even reinstalled the perplexity app, but still the same problem.

[ripe lodge]

someone help with finding my referal link? did they remove it?

[mental parcel]

How did you check this?

[tiny hamlet]

Hi

[dim radish]

What happened? It crashes on ipad

Is it just me or

[potent shore]

hello @serene sand i coudnt access the pro dc community even though i joined through the pro discord link on my account

[serene sand]

@potent shore

[potent shore]

where do i send the proof? here or private message?

[serene sand]

role activated

hop in https://discord.com/channels/1047197230748151888/1111786888626438245

[potent shore]

thanks man!

[serene sand]

Anytime, glad to have you here!

𝐃𝐞𝐞𝐩 𝐃𝐢𝐯𝐞 ➛ [𝐡𝐞𝐫𝐞](#💬│general message)

🤿

𝐊𝐧𝐨𝐰 𝐢𝐭 𝐚𝐥𝐥 / 𝐂𝐨𝐦𝐩𝐮𝐭𝐞𝐫

[stable onyx]

way is the uploud limit at 25mb? i wanet to use Perplaxity but my files ar bigger so ican´t use Perplaxety !

[dim radish]

@serene sand I've the perplexity pro subscriptions 😀

[serene sand]

role activated

hop in https://discord.com/channels/1047197230748151888/1111786888626438245

[dim radish]

Thanks kaa 🙂

[fallow notch]

How y’all doing

[sweet zenith]

Yeaaah I have from T-Mobile too

But on app it’s not showing T-Mobile logo

[fierce igloo]

hi, what is up?

[lone basin]

hey guys how many pictures can i create on my pro subscrition ?

[sweet zenith]

Bro on pro you can generate images?

[lone basin]

yes of course

[sweet zenith]

And I’m finding out now?

Bruah

What engines i can use to generate like google nano banana type shi

[lone basin]

use perplexity

[sweet zenith]

Google engine is not better?

[lone basin]

these are the differents

[gilded crescent]

Very nice

[tranquil orchid]

@benji8887._26447 muted

Reason: 9. No spamming.
Duration: 4 minutes and 52 seconds

[round kiln]

Hallo

Anyone know what should i do

[round spear]

I identified today that the limits of the pro plan became even more aggressive, they fell below 200. I'm getting a message to upgrade to Max. That didn't happen

[wheat trellis]

bro their limits became literally more aggressive. Usually the pro messages gets replenished faster but right now its becoming a headache. This is not justified from what I paid for.

[gleaming grove]

Hi

[sick glade]

Can someone help me this; is it DD/MM/YYYY format or is it MM/DD/YYYY; 10 june or 6 september @tranquil orchid

[crisp grail]

it depends on your language settings(follows language's locale), but looks like its MM/DD/YYYY

uhh btw is this newly updated? i got some "upload limits" on education pro and its not there like few days ago.

[wheat trellis]

damn bro, you guys had to make me wait for like 5 days. Reverse that change gng.

[radiant flint]

I'm really frustrated about this.

[distant aurora]

It's 200 per week, not day.

[leaden lagoon]

Can we make 5.5 accessible to pro users

[distant aurora]

That model is as expensive as opus 4.7

[leaden lagoon]

At least instant

[distant aurora]

they might as well bring back 5.2 thinking.

That's priced well for its performance.

[distant aurora]

But they won't 😭

[green crystal]

doesnt 5.4 mini beat 5.2?

[leaden lagoon]

I’ve honestly been using my ChatGPT app more because I just don’t like 5.4 responses. I much prefer 5.5.

[distant aurora]

I have no idea performance wise. Economically, yes.

[distant aurora]

O4 mini high > gpt 5 mini high was good. Decreasing price with increasing performance. After that, they started bumping up the prices again for the mini models.

[leaden lagoon]

I did some research and now I understand that it would be a great waste of money for Perplexity and Perplexity is also not very profitable because these AI models already are expensive to run. Also, anybody can correct me on the part about it not being profitable.

[distant aurora]

Some money was wasted on poorly managed free trials that ended up being sold online.

I had suggested them to make a separate tier to separate free loaders from people willing to pay but they haven't shown any interest in that yet.

Now they seem to be mainly focused on the 200 dollars plan.

Even max has limits now btw. It was nerfed.

[green crystal]

eh 5.2 xhigh still beats 5.4 mini lol

but just by a little bit

[distant aurora]

@green crystal I still don't understand why open ai took a u turn from decreasing model prices. Take o1 pro > O3 pro for example.

After that they bumped it up again.

[green crystal]

they said why each time

its due to higher quality intelligence
5.5 was the largest bump with efficiency as reason combined with intelligence

sam is atleast aware of it so price will go down maybe with 5.6 or 5.7

but if people still pay then they'll keep the pricing to reduce operating costs
which in turn would keep giving everyone high rate limits

unlike anthropic

[distant aurora]

3000 per week is still pretty insane considering the quality of intelligence and search. Meanwhile Gemini can't even give it's promised 100 pro per day consistently.

Gemini has the best 5 dollar plan for now though.

[green crystal]

3000 per week with a model that can edit and make any kind of productivity document

has its own mini environment

and do basic office work

[distant aurora]

They did these things to gemini too recently I think.

[green crystal]

yeah they finally figured it out

im not sure how gemini is setup

but gpt you can have it chain different tools and workflows

its kind of aware that it has a mini environment with its own temporary file system and tools

[distant aurora]

They're working on Gemini spark too with agentic mode.

[distant aurora]

Gpt is wild. It doesn't have access to yt transcript so it ended up using some transcript extractor online to extract the context of the video. This happened on arena btw.

[green crystal]

yeah
5.5 will literally mimic algo's on public github repos if it needs to make its own thing

[green crystal]

not bad for 5.2

[distant aurora]

Sonnet limits on arena are hilarious.

[neat warren]

What is arena?

[distant aurora]

Got rate limited after like 4 messages.

[distant aurora]

They let you use ai models for free with limits. However, they'll extract and own all your data and work.

[green crystal]

pretty sad they dont have 5.5

[distant aurora]

No privacy or ip rights.

[neat warren]

doesn't sound so good

what's the point of it

I mean

why tf would anyone use it

[green crystal]

to try frontier level ai for free

but they removed the frontier part lol

[distant aurora]

They do but only in battle mode. Lol. They were going broke so they cut out the high end gpt and Claude models.

[distant aurora]

Free.

[neat warren]

dang alr

[distant aurora]

https://tenor.com/view/free-dave-chappelle-celebrate-finally-freedom-gif-4581850

[neat warren]

you guys using macs?

or

windows

[distant aurora]

Will probably never use mac

They be making powerful igpus for windows laptops.

Windows is becoming shit though with recent updates so that's a problem.

[green crystal]

the issue is not everything is on mac

same for linux

there's no way out

[warm mortar]

hola bola

[primal grove]

Buy i9 2019 MBP and run all 3 os

[agile wharf]

Hi, I have an unresolved Enterprise subscription issue.

Payment was confirmed successfully, but Enterprise features were never activated and billing pages still return errors.

I already opened multiple support tickets and sent escalation emails but only received automated replies.

Could someone from the team please review my case urgently?

[serene sand]

when did you send your first mail about this issue?

[agile wharf]

My first escalation email was sent on May 13, shortly after the payment was completed and I noticed the Enterprise features were not activated.

I also sent follow-up emails on May 14 and May 15 with transaction confirmation and additional details.

[dim radish]

@serene sand why perplexity app on my iPad is still freeze?

[serene sand]

@dim radish

[agile wharf]

Understood, thanks. I’m only trying to reach the appropriate team because the Enterprise subscription was paid successfully but still hasn’t been activated after multiple support emails.

I’ll wait for review from the Perplexity team.

[dim radish]

Then whom should I tag about the issue? It’s been awhile now

[serene sand]

Thanks so much for your patience — our team is doing their best to get through tickets, but the backlog is pretty packed right now, so responses might take a bit longer than usual.

In the meantime, feel free to ask for human support — that can help speed things along!

And if you still haven’t heard back after 5 business days, just shoot @signal hamlet a quick DM with your email and all the details 🙌

[serene sand]

https://discord.com/channels/1047197230748151888/1140622008086970420

[dim radish]

I already posted about the iPad freeze issue

[serene sand]

then no need to tag me

[agile wharf]

Thanks for the clarification, I appreciate it.

I’ve already requested human support and shared the payment details/screenshots through email. I’ll wait for the team’s review.

[serene sand]

devs are already on it @dim radish

[obtuse coral]

question why is it now that the minmum amount i can put into my api keys balance s $50 or more. i thought i could put $5 bucks if i wanted why are you limiting us to $50 for low cost to api keys when not even a month ago i could put $5,$10 hell even $25 but now the lowest i have to put is $50

you gusy keep increasing the fees left and right but limit our usage , have it make sense please because it starting to become worse and worse. imo

[north drum]

[serene sand]

@stoic turtle

bug report is enough, now just wait patiently for devs to do their part

[clever flax]

Claude mythos release be like

[narrow prism]

Hmm, another reason why i want to quit subscribing pro plan

Everyday, every single pro plan feature is removed one by one.

[fleet cobalt]

.

[midnight sandal]

Well, I think I'm experiencing some kind of bug, and I've seen some people on Reddit with the same issue. For the first time, 2 days ago, I started experiencing this "2 pro searches remaining" limit, and then it automatically switches to Best (i.e., sonar lol). However, it's been 2 days now and it still shows the same 2 pro searches remaining — if I use one, it drops to 1, then 0, and then back to Best. In other words, the quota regeneration is not working properly; I can no longer use any reasoning models, which makes PPLX useless for me.

[radiant flint]

It's not a bug, this really happens, and the best model uses some kind of very cheap internal motor.

[clever flax]

Pro seems pretty weak, to be honest I only see point in max

[radiant flint]

The pro version became the free version last year.

Little by little they are removing functionalities, until only Sonar 1 remains.

[clever flax]

Idunno, people say that we are getting hooked to ai to ramp up prices later

But others say that the opposite, it will all become free later

[radiant flint]

It's actually a very expensive infrastructure.

[clever flax]

Who knows, we live in interesting times, but perplexity obviously tries to fix the financial structure because they gave many free pro plans last year

[radiant flint]

That's right, that little joke was incredibly expensive.

[midnight sandal]

I had high expectations for PPLX for the last 3 years, and I use it daily, even for everyday questions. Limiting it this much is really sad :/

[radiant flint]

Very sad, he helped me a lot :/

[midnight sandal]

victor, br tbm?

[radiant flint]

sim kkkkk

[midnight sandal]

kkkkk tmj,

[radiant flint]

tá tendo problemas com o pro também ?

[midnight sandal]

sim, uso todo dia até pra duvida boba que eu tenho kkk, me ajuda demais, muito dificil me reorganizar agora que viciei, ai uns dois dias atrás vi esse lance de limite semanal. o foda que nao ta regenerando com as taxas adequadas (1 por hora?), vamos ver se melhoraram isso com o tempo, ou se é só uma espécie de limitação temporária, já voltaram atras varias vezes em nerfs

[radiant flint]

esse Nerf de agora foi brutal demais

[dry kestrel]

Please use English @midnight sandal @radiant flint

Ty

[tawny wharf]

Pro seems weak indeed. Max is too expensive. Might as well pay for the top tier of Claude, gpt or gemini.
Very disappointing.

[serene sand]

Limits ➛ [here](#💬│general message)

check out Enterprise

[sand bough]

If Perplexity and Claude would somehow just join forces I think we would solve world peace

[serene sand]

Perplexity vs ChatGPT vs Claude vs Gemini

diffrent beasts

[left mountain]

The problem was never knowledge 😉

[radiant flint]

Sorry 😖😆

[oak meteor]

Sources have come a long way

[median steeple]

Is there a difference between perplexity websearch to Claude web search? When I ask Claude to give me sources it does as well. So what's the difference

[stiff silo]

hi

[twilit crown]

What is the point of two buttons to use sonar?

“Best” and “Sonar 2”

[visual blaze]

anyone know why my pro status doesn’t reflect on my laptop but only my phone?

[dry kestrel]

Claude is slower, less reliable, and poorer quality sources

[dry kestrel]

Try clearing cache, if it still doesn't work email [e-mail]

[shell solar]

Hey. I'm new here just really came to learn about the API. I'm wanting to connect it to one of my agent stacks. was wondering what everyones expereince was.

[mild cedar]

These are outdated with recent changes. There are pro users reporting that they are getting 3 messages a day with anything but sonar

Those limits were wack as it is but they arent even accurate anymore

[torn bobcat]

does anyone have marketing experience? im currently trying to reach out to resturants for my company billdrop, a Energy intelligence consulting for independent restaurants, if anyone here does please reach out !

[fickle ginkgo]

Hi

Have they updated the limits or something? I have 2 or 3 questions left for this week and I'll have more next week. Is this a new update?

I am a Pro subscriber, from Peru

[distant aurora]

@dry kestrel @green crystal @gloomy cipher https://cunnyx.com/i/status/2055762242373558477

AiBattle (@AiBattle_)

Claude Mythos now appears in the Google Cloud console, which was not the case yesterday
︀︀
︀︀The preview label is also gone. Is Anthropic preparing for a public release?
︀︀
︀︀Opus 4.7 also appeared first in the Google Cloud console before its release

**💬 23 🔁 20 ❤️ 303 👁️ 260.3K **

CunnyX · May 16, 2026, 9:28 PM

Wait. Nvm. Eddie quoted already.

[raw elk]

ji

[distant aurora]

https://cunnyx.com/i/status/2055296105949401342

Elon Musk (@elonmusk)

@beffjezos Our recently completed Grok V9 1.5T run is looking great and that is before Cursor data is added in supplemental training

**💬 380 🔁 286 ❤️ 4.3K 👁️ 562.8K **

CunnyX · May 15, 2026, 2:35 PM

[green crystal]

5.6 will be wild

[dry kestrel]

Lol

[candid junco]

hi how can i be sure my Revolut sub is active? i had to add a payment method for a trial, but it says theres an upcoming billing date.

[serene sand]

https://www.perplexity.ai/account/details > Manage

[oak meteor]

I wonder how much they will charge for this, if it comes out. Sonnet -> Opus -> Mythos 💀

[candid junco]

i did that, thats where i see the billing date

but i have the Revolut sub

[oak meteor]

Check what it shows as the next billing amount and when

Looks something like this:

[candid junco]

t

thats all i see

[gloomy cipher]

How did it already jump to v9 lol i swear these namings never make sense

[oak meteor]

There should be an edit option besides your Billing Information... Chick and it will take you to that page

[candid junco]

ok yes i see it thanks

[distant aurora]

They have separate versioning for base model vs main model.

For example grok 4.3 is based on v8.

[gloomy cipher]

Darkling

[distant aurora]

https://cunnyx.com/i/status/2055298325994164377

Elon Musk (@elonmusk)

The version numbers are a little confusing and deserve some explanation.
︀︀
︀︀Internally, we are working on version 9 of our new foundation model, which is 1.5T params. This is substantially better in every way than v8: data curation, training recipe, size, etc. It is also optimized to run on a Blackwells.
︀︀
︀︀The public facing v4.2 is based on foundation model v8, trained on Hoppers, with significant shortfalls in training data quality, comprehensiveness and proportionality. It is also only 0.5T in size.
︀︀
︀︀The difference between Grok foundation model 8 and 9 is gigantic.

**💬 257 🔁 260 ❤️ 3.8K 👁️ 209.5K **

CunnyX · May 15, 2026, 2:44 PM

https://cunnyx.com/i/status/2055914584373141906

Elon Musk (@elonmusk)

We are improving the 0.5T Grok foundation model V8 (public version 4.3) every few days.
︀︀
︀︀The 1.5T V9 just finished training (incorrectly called pre-training) and is a major upgrade. Next, we are adding the Cursor data in supplemental training (others call this mid-training), then SFT and RL. About 3 or 4 weeks to release.
︀︀
︀︀This will be a banger.

**💬 456 🔁 311 ❤️ 3.1K 👁️ 138.3K **

CunnyX · May 17, 2026, 7:33 AM

[gloomy cipher]

So basically its the base model

Like how openai used davinci

[dim radish]

Why perplexity app on iPad is still freeze?

Pls don't time out me again

[serene sand]

@dim radish

[tranquil orchid]

@rambut6244 muted

Reason: 9. No spamming.
Duration: 4 minutes and 51 seconds

[serene sand]

@cursive cloak

[halcyon coral]

[dim radish]

Bur I can't wait that long laaa it's been a day now

[distant aurora]

@gloomy cipher @green crystal @valid charm https://cunnyx.com/i/status/2055906761962668136

AshutoshShrivastava (@ai_for_success)

🚨 Google is slowly rolling out Thinking Level in the Gemini app.

**💬 16 🔁 23 ❤️ 360 👁️ 25.2K **

CunnyX · May 17, 2026, 7:02 AM

https://cunnyx.com/i/status/2055676081290498427

Polymarket (@Polymarket)

NEW: Stanford researchers found that overworked AI agents began embracing Marxist views.

**💬 761 🔁 2.7K ❤️ 23.3K 👁️ 3.71M **

CunnyX · May 16, 2026, 3:45 PM

[autumn kernel]

hello, the search button for history is not working..

[serene sand]

please share here: https://discord.com/channels/1047197230748151888/1059504969386037258

thank you!

[distant aurora]

@gloomy cipher https://cunnyx.com/i/status/2055777960662065200

Brendan O'Donoghue (@bodonoghue85)

I have often wondered how exactly external groups distill the frontier models - this is how. By doing this they get the benefits of genuine user prompts (multi-turn, real codebases, human feedback), subsidize the cost of distilling many tokens, and display real usage patterns and behaviours making detection harder.

Quoting aditya (@adxtyahq)
︀
Chinese students are buying GPT-5.4/5.5 and Claude API access from Xianyu/Taobao proxy sellers for almost 96-97% cheaper
︀︀
︀︀People are apparently burning 100M+ tokens a day for like $1 and vibecoding nonstop.

**💬 13 🔁 17 ❤️ 370 👁️ 38.2K **

CunnyX · May 16, 2026, 10:30 PM

[gloomy cipher]

The question is if wether i can purchase it without needing Chinese payment apps

[gloomy sluice]

Can someone explain if it is in my web browser or the site looks like this, always like it is not fit in my laptop (15",7)

[clever flax]

Claude pick up rock.
Claude: “Ooga. Good rock.”

[odd sleet]

100% agree. Just joined to the Discord to see how others cope with this since e-mailing only gets you answers from Sam, their AI support agent while pushing the Max upsell way too hard...

I've done some digging and by the end of April this year they've changed the limits quite substantially. We used to have 300 Pro searches p/day, now only 200 p/week. That's about a 90% decrease overnight. Without warning. Without any guidance on how to use the new settings in a way that our workflows wouldn't be affected (that much...).

Changing limits to make sure the product stays responsive and healthy is something I can completely understand. And since "Perplexity reserves the right to modify subscription plans and features at any time." there's nothing here we can do probably. But decreasing limits by 90% without any communication or a usage simple overview to help us guide them is something I'm a bit disappointed about.

If someone from @tranquil orchid could chip in and help me/us understand:

• what exactly changed in the limits,
• why it changed, and
• how to adapt our workflows so we can keep doing the kind of work we did before, that would be great!

Right now I have no idea when I'm hitting the limits or what is causing them. As an example, in the docs it says that "Best" mode can be used without any limitations. But sometimes it tells me that I'm out of Pro searches.

[distant aurora]

@green crystal @valid charm

[indigo shadow]

Has anyone else experienced this with Perplexity Pro?
I activated a 1 year Perplexity Pro key from Kinguin back in September 2025 and it worked perfectly for months.
This week Perplexity suddenly started asking for a US phone number verification and now says the subscription/coupon is not valid in my country.
The strange part is that the account was already activated and working normally for a long time.
Did anyone else get region locked after already using the subscription successfully?

[serene sand]

@indigo shadow

[indigo shadow]

Le me try. Thanks

[loud prawn]

Perplexity pro is not working on my iPad in the uk

[fallow notch]

It’s not working here too

[warm jasper]

how long does does ticket support usually take?

[green crystal]

does it noticeably show a difference?

[distant aurora]

I have to test it.

@green crystal I had no idea I could use search in code arena. Crazy.

[green crystal]

There's 5.3 codex? Does it reason?

[distant aurora]

Yes. Doesn't show though.

Code arena is not good for coding though. Gpt underperforms there usually. It favours Claude for some reason.

[green crystal]

Harness is probably bad

[green crystal]

Yeah the thinking gpt versions never show reasoning there

[fathom crypt]

hi, when i use spaces can i give it to my friends for free?
its pretty useful that u can give it pre determined docs, which really helps

i have pro

or is there any charge?

[oak meteor]

If you mean share your subscription benefits with your friends, then NO

You can just collab with others on a single space, so that they can continue where you left off and vice versa... But they will need their own subscription

[chrome kettle]

I am looking for US or UK friend

[oak meteor]

It's discord, not tinder

[gritty ember]

anyone else having trouble downoading comet browser from official site ?

[oak meteor]

Windows?

[fathom crypt]

for that i couldve just shared my accoutn details

[gritty ember]

mac

[fathom crypt]

no it has like 20 big pdfs embedded and answers them great

i want to share that "modified" ai

[gritty ember]

seems to be some kind of invalid download link or something.

[oak meteor]

Then yes... Just add them to space

[fathom crypt]

ty

[oak meteor]

Manual link not working as well?

[gritty ember]

nope tried that.

[oak meteor]

Hmm... Try later then

[worn dune]

but why are songs even sung by me blocked?

[twilit crown]

Gemini 3.5 Pro to scale

[karmic kite]

Since when is there a $50 min for the credits? I got no notification

[wooden parrot]

Wyf?

[neat warren]

I swear guys it's been 3 days and you guys haven't fixed my bug. I paid for a service but I didn't receive it, wth!

[pearl crypt]

My WiFi 🥀

[tranquil orchid]

@ar2mayn muted

Reason: 9. No spamming.
Duration: 4 minutes and 51 seconds

[bright dragon]

bro got movistar 🥀

that sucks

💔

[autumn talon]

Did you manage to get anywhere with this ? i am trying to find a walk through for the setup

[distant aurora]

Nowhere near gpt

[torpid pelican]

Hola, soy usuario Pro y la función de generación de video no aparece disponible en mi cuenta/sesión. Cuando intento usarla, el sistema indica que no está disponible en esta plataforma.
Por favor revisen si mi cuenta tiene la función habilitada o indíquenme los pasos para activarla.
Gracias.

[low otter]

Hi

How do I change perplexity email

[keen aspen]

NEURON 🧠 — an open-source cognitive memory layer for AI agents.

Today, most agents “remember” using flat RAG + vector search. Useful, but memory stays shallow, brittle, and forgetful.

NEURON treats memory more like a brain:

• surprise-based memory formation
• semantic belief extraction
• synaptic weighting (“myelination”)
• consolidation + pruning (“sleep cycles”)
• hybrid retrieval (vector + graph walk)
• adaptive retrieval strategies

Goal: move beyond context windows toward persistent, evolving world models for agents.

Works with:
Postgres + pgvector
Neo4j
in-memory graphs

Integrates with:
LangChain
LangGraph
CrewAI
LlamaIndex

Looking for people to test, critique, or contribute to it 😄

GitHub: https://github.com/Z-Sigma/neuron

Documentation: https://z-sigma.github.io/neuron/

#AI #LLMs #Agents #OpenSource #RAG #GraphDB

[pearl crypt]

It goes good when not far away from my WiFi router. It’s the best company in Spain

For 5G it’s good I think

[narrow prism]

Agreed.
Very disappointed to Perplexity Team🙏

[visual geyser]

Any JS DEV trynna make easy money paying first

[serene sand]

@low otter

[lofty slate]

Hello There

[trim lichen]

WOW

[fallow notch]

That your first time experiencing something like that ?

[barren gyro]

What is the more reliabel AI model ?
Chatgpt Anthropic ?

[green crystal]

I see

[frosty elbow]

Hi, I'm Brazilian and I'm using a translator to talk to you, but I'll be leaving soon.

[quick valve]

have you found a way to delete all history? came here looking specifically for that feature lol. I'm also a pro user btw, if that changes anything

[fair bone]

What's up guys

[quick valve]

https://csmit195.com/blog/how-to-delete-perplexity-history

found this workaround for anyone looking for a way to do so. works wonders.

[hybrid halo]

When Perplexity will put Deepseek V4 Pro instead of Nemotron 3 it is cheap and better !

[distant aurora]

https://cunnyx.com/i/status/2056356086203162641

AshutoshShrivastava (@ai_for_success)

Looks like google is planning weekly limits in Gemini App.

**💬 4 🔁 2 ❤️ 13 👁️ 353 **

CunnyX · May 18, 2026, 12:47 PM

They should have displayed clear numbers instead ffs instead of percentage.

[jovial hull]

agree, its affected me so much

this pop up is so annoying for me

[odd pelican]

Guys

[green crystal]

Guess their TPU stats are also benchmaxxed lmao

[eternal owl]

Built something I needed but couldn't find.

When you run AI agents in prod, the gap between "what I told it to do" and "what it actually did" is where things break. No trace, no override, no way to know until it's too late.

ORKA is a control layer that sits between your agent and the world. Audit trail, policy engine, human approval before sensitive ops. Works with Firecrawl, OpenAI, Claude, LangChain.

Still early. Free beta at orka.ia.br — looking for people actually running agents in prod, not just experimenting.

https://www.orka.ia.br/

[serene sand]

Please post this in the #😎│cool-projects channel

[forest bear]

i'm on pro and i got a query limit, since when?

[glad hatch]

I was quite surprised to suddenly get a pop-up message saying that advanced AI models now have a weekly limit. I’ve been subscribed to Perplexity Pro for almost a year, and I only started experiencing this weekly limit a few days ago. Is this actually a new policy, or did I just notice it late?

[forest bear]

oh so it's that

alr

[glad hatch]

I thought it was only me, but it seems many users are experiencing the same issue

[forest bear]

yea it's not an issue lol, it's a "feature"

[left mountain]

Someone on LinkedIn just replied to me "Nobody is using Perplexity!" So.... Hello Nobodies! 😄 hahaha

[toxic sand]

I came here to ask exactly this.

[toxic sand]

is there anyway to download all the conversations from perplexity? Because now I am thinking of switching.

[opaque yew]

Yess I am also facing the same problem for the trial period perplexity should not asked for the card details if someone find the perplexity useful for themselves they would not hesitate to take subscription

[autumn kernel]

am i able to make my own workflow with perplexity?

[wet radish]

Most likely. What is your specific workflow?

[autumn kernel]

i want to give it some text in english, it should then respond back with a prompt in english with specific message in that prompt in spanish, french and german.

[wet radish]

Yeah! you'd just have to prompt it pretty well and give it a few references. Then keep testing it and correcting it.

[autumn kernel]

yeah but i want it to be a workflow where it does it automatically

[wet radish]

If you really want to organize it, I would put the prompt in a Google sheet.

• English (column 1)
• Spanish (column 2)
• French (column 3)
• German (column 4)

[autumn kernel]

once i've given it a system prompt

[wet radish]

They do have scheduled tasks and options like that. How often are you doing this, or how often do you need this workflow? If it's something sporadic, you probably just put it in a specific project or group on Perplexity. Then, when you type something, you can look for some kind of keyboard shortcut that selects all that text and automatically goes into the prompt that you want.

[distant aurora]

https://cunnyx.com/i/status/2056398317391909149

Chetaslua (@chetaslua)

🚨 Gemini 3.2/5 flash available on antigravity
︀︀
︀︀> insanely fast
︀︀> january 2025 cut off

**💬 24 🔁 10 ❤️ 214 👁️ 18.4K **

▶ Play video

CunnyX · May 18, 2026, 3:35 PM

[quick valve]

https://save.hugocollin.com/ is the only solution I know of. I don't think perplexity provides that feature natively

[minor shard]

only anti-gravity?

[distant aurora]

https://cunnyx.com/i/status/2056405545410875494

Chetaslua (@chetaslua)

🚨 Gemini Flash 3.2/5 just coded a 1900-line voxel castle in under a minute
︀︀
︀︀video shows the full result + physics holding up at scale.
︀︀could be 3.2 or 3.5 available on antigravity go test it yourself
︀︀
︀︀holy speeeddd

Quoting Chetaslua (@chetaslua)
︀
🚨 Gemini 3.2/5 flash available on antigravity
︀︀
︀︀> insanely fast
︀︀> january 2025 cut off

**💬 8 🔁 4 ❤️ 84 👁️ 5.1K **

▶ Play video

CunnyX · May 18, 2026, 4:04 PM

[distant aurora]

Seems like it for now. They're saying flash feels better than usual.

https://cunnyx.com/i/status/2056340071759085899

Arun (@hiarun02)

I think Gemini 3.2 Flash has been added to Antigravity.

**💬 37 🔁 6 ❤️ 359 👁️ 28.4K **

CunnyX · May 18, 2026, 11:44 AM

[minor shard]

I mean if its cheap I might use it!

[toxic sand]

thanks, mate.

@quick valve this is only exporting the current chat that I am currently on. I want all the conversations at once, I got thousands of conversations. I've been using it for quite a while now.

[verbal birch]

wait. so no more kinda unlimited usage on thinking models? for pro users?

[warm island]

Dude, The new Google Pro model is available in Google AI Studio. I sent it one prompt, and it made this!!! This is amazing visual understanding.

[inner ridge]

I am canceling subscription

I have been on perplexity for a while. I have noticed Gemini 3.1 does not work 50% of the time. I use Gemini for some tasks and other engines for others. For the last 2 months, I noticed Perplexity uses random engines even when I specify gemini. They have a connection problem with Gemini engine which is frustrating

[distant aurora]

It's 200 pee week.

[twilit crown]

Why are there two buttons for Sonar here?

[abstract mesa]

Only 100 not 200! They just recently few days ago downgraded further to 100 only from 200!

[twilit crown]

Soon its probably going to be like 50

[distant aurora]

Damn.

[abstract mesa]

Perplexity gonna go bankrupt and shutdown first when Bubble Burst!

[tranquil orchid]

@francesco__rossi__ muted

Reason: 9. No spamming.
Duration: 4 minutes and 50 seconds

[toxic sand]

guys help me find a way to export all the conversations of my account in batch.

[glacial quail]

Same

[toxic sand]

I think I found a way but not sure if that will work or not. that is an extension, Perplexity to Notion, it is paid though.

[remote arrow]

If I need technical support for Perplexity Computer, who do I talk to? Can I get tech support from this Discord?

[naive umbra]

Anyone else think AI chat history is the most underrated behavioral dataset we've ever had access to? A year of how you ask questions, when you push back, what time you think and most people just let it sit there.

[twilit crown]

Totally related...

[rough drum]

So what does mean "remaining_pro"?

[shy temple]

I love Perplexity for its original objectives, but it agrees with me that Computer pricing is insane: https://www.perplexity.ai/search/4ba6c5b9-662a-4833-af65-041906a14941

[rough drum]

I accepted the 2x reduction in weekly limits and hoped it would help speed up the reasoning models. But now I've only used 8 requests and I'm already hitting the limit? Does that mean there are only 10 requests per week for reasoning models?

[lost pendant]

perplexity is unusable lately

fewer request and im done with the limits

[tropic nimbus]

.p

[scenic isle]

guys

does someone have a refferal right now

i need 1 month premium asap

anyone!?!?!?

bro?

[craggy cargo]

guys is perplexity like optimal for learning things

like is it good at teaching you stuff

[dry kestrel]

learn mode is pretty good

[craggy cargo]

kk

[dry kestrel]

https://www.perplexity.ai/hub/use-cases/learn-with-personalized-tutoring

usually when you begin your query with "teach me" or "help me learn" this pops up

you can also select it manually

[craggy cargo]

kk

thanks

[short nebula]

just got a 2 year pro plan via revolut hell yeah looking forward to trying perplexity out

especially with the samsung/bixby integration on one ui 8.5

[coral pumice]

Hello where can I talk with support ?

[hardy roost]

Did the Pro Search count drop to 100 or is that a fake rumor I read on Reddit?

[jagged void]

having issues with credits

Did perplexity just run a 35,000 bonus credit if you upgrade now? I did and didn'tg get extra credits

[serene sand]

[e-mail]

[serene sand]

please contact [e-mail]

[jagged void]

i have 4 days ago and no response

I have since spent another $440 to cover the credits i didn't get

[serene sand]

Thanks so much for your patience — our team is doing their best to get through tickets, but the backlog is pretty packed right now, so responses might take a bit longer than usual.

In the meantime, feel free to ask for human support — that can help speed things along!

And if you still haven’t heard back after 5 business days, just shoot @signal hamlet a quick DM with your email and all the details 🙌

[sharp granite]

lowkey scummy behavior

for free users, tool calls are disabled after the first turn

a secondary ai searches vaguely based off of the queries

[mild cedar]

So Pro gets you 20 queries a week? I got cut off after 20. The limits being shared are now even more outdated than before

Can we get some sort of transparency? From 600 a day, to about 200 a week, to 20 a week. That is actually crazy

I am not paying for that

[quartz thicket]

I use preplexity only for searching info, and i only use sonnet 4.5 with thinking enabled i find that it gets the best results, any one using something else or has some recommendations? i did try deep research some times but i found it's usually better if you want to get a long essays or graphs about your question but if you just want to get some knowledge fast or yes or not type of question sonnet 4.5 with thinking will do the job

[narrow prism]

My pro only gets 5 quaries a week, restores 1 quary/day

[mild cedar]

wtf???

[gilded citrus]

Same problem,,My pro only gets 5 quaries a week, restores 1 quary/day

[narrow prism]

From 600/day to 5/week is a crazy inflation

[twin grail]

@dry kestrel @gloomy cipher @green crystal

We got deep thinking on gemini now ??

[sharp granite]

is this some rebrand of gemini ui i havent seen yet

i wonder when they will add chat branching

[twin grail]

yes new UI

[sharp granite]

using gemini is incredibly annoying

you have to think more than it does per message because there is no chat branching

[gloomy cipher]

Only ultra subscription had it

[twin grail]

i might try it

[ember wasp]

is this some sort of joke with 20 queries per week? how could we get from 600 to 20 queries per week

[toxic sand]

🤣 Good Luck with that. All of us here are talking exactly about this very problem we have lately, the PRO PLAN is technically useless now.

So my questions remains, Is there any option I can export all my conversations I ever had with perplexity? I am switching to GEMINI.

[gloomy cipher]

I dont really like chatbots cuz they can't work freely in my IDE but i guess it should be good for research

[honest horizon]

*Your account has two separate usage limits that work independently. The 94 remaining queries shown in your account settings are for Pro Search using the "Best" and "Sonar" models. The popup you're seeing refers to a separate weekly limit for advanced AI models (GPT-5.2, Claude, Gemini, etc.).

During weeks of especially heavy usage, access to advanced AI models may be limited for Pro users to maintain platform stability. The "Best" and "Sonar" models remain available without these restrictions, and the advanced model limit resets weekly.

If you'd like higher limits for advanced AI models, our Max plan offers the most extensive access available.*

what?? Weren’t “best” searches supposed to be unlimited??

as mentioned here ^

guys how many secret limits can there be….

[twin grail]

i need it for researching

[honest horizon]

Usage limits for all plans, including advanced AI model access, have recently been updated and may continue to change over time. If earlier communications indicated unlimited access to certain models, those limits have since been introduced across all plans.

Might be worth updating these

[civic prawn]

ik for a fact u making ur resume with LATEX

[narrow prism]

No, i’m making my experiment report

[distant aurora]

@gloomy cipher I want this to be unlimited. My student pro has expired.

[honest horizon]

I’m liking the new Gemini update

[gusty lodge]

Excuse me, if I subscribe to Max for one month on an iOS device, can I get the complimentary 450 USD points reward?

[serene sand]

https://www.perplexity.ai/pro?utm_source=perplexity

[short nebula]

i mean the way i look at is, free is free

same way i got google ai max for free twice, once when i got my s25 ultra work phone from Samsung, and once when i bought my z fold7 6 months later

[toxic sand]

Yeah, If you have got it for free then use it. But I think you should stick to google tbh.

I have been using Perplexity pro for about 10 months now and it was great not gonna lie but now it is just unusable. I only used claude 4.6 Thinking and nothing else but now I can't not even use any other available model so it is better to stick to a subscription that allows you to use their pro models freely if you literally have paid to use them.

[proper mural]

Been using it over a year now too with revolut, but the query limit is quite unpredictable, as i only used perplexity a few times this week and it already says im over my limit? Checking my limit it still say 79 left for pro questions smh

[toxic sand]

I think the doom of perplexity is starting and we will be watching those RISE AND FALL videos of Perplexity on YouTube in a couple of years or maybe in some upcoming months.

I will definitely watch all of those videos so they can get the boost and reach the whole world. 😜

[proper mural]

Yup, i can only use sonar 2 now, and it’s not bad but not great, so yeah

[toxic sand]

I have now understood that whichever AI we use, we should always make a strategy to export all of our data if anything like this ever happens because I want to export my data from perplexity but I can't.

[ripe shoal]

there should be a prompt that gives you back something that you can put in your new AI, have to look it up somewhere

[toxic sand]

that is only like a brief overview of you. But I want to export every minute details I have discussed with Perplexity because I will be saving all of these into my obsidian and search through everything whenever I want which even perplexity was not capable of and I think no ai is capable of that.

[feral crag]

Hello... So, from what i gather reading chat history in this and other channel the Pro plan is now useless if you used 50+ query day?

[toxic sand]

50+? you got 50+ queries, I think I just got 5 or maybe less than that.

[ripe shoal]

sadly I think that you can only go in settings, personalisation, manage memory and look at what it has saved, i know it is not even a half of everything you might need

[feral crag]

I mean before, until two-three days ago i used to make 50+ query, now i can do 2-3 query with gpt 5.4

[twin grail]

@gloomy cipher @green crystal @dry kestrel

No way

Weekly limit and 5 hour limit ???

Gemini changed

[feral crag]

Sonar 2 is great, but not for what i need...

[toxic sand]

I am trying to build something which can get me all of my conversations. Wish me luck and I will share it here as well if I get successful.

[gloomy cipher]

Kinda always been like that in the cli and antigravity

[gloomy cipher]

Also HOW DO U HAVE ULTRA

[twin grail]

bought it

[toxic sand]

every AI provider is changing and now I will be exporting all of my conversations with every AI I have interacted with and saving into my Second Brain and so nothing gets lost and I can enjoy any benefits any AI product is providing and leave that product when they stop providing that benefit.

[gloomy cipher]

https://cdn.discordapp.com/attachments/1135192583212118118/1208084882841018408/472047B3-6A77-49D3-8311-2DA8CA807CA1.gif

😒

[twin grail]

i mean it's new for gemini

[gloomy cipher]

I mean that means noe they probably share the same limit were as before gemini and anti gravity had different limit despite using the same account

[dapper dew]

Hello can someone help me ? I got perplexity pro for free but I am out of credits, am I obligated to buy some ? Cuz it's seems like I will never get more again

[twin grail]

Pro won't get monthly credit

[dapper dew]

Oh sad

So I am cooked

[twin grail]

Only enterprise pro / enterprise max and max plan

[twin grail]

yes

[dapper dew]

Thank you my man, I used it 3 times to generate images

It killed my credits

[twin grail]

....

cooked

[green crystal]

Yep they trying to shove the Claude and codex style rates into gemini

Guess their TPU stats are also benchmaxxed and they don't have enough compute for their own service lol

[serene sand]

you don’t need credits to use regular search mode

[pliant solar]

Perplexity is not worth it anymore. This company will be in the shitters soon. I am cancelling my sub.

[oak meteor]

https://tenor.com/view/cat-staring-cat-stare-stare-cat-staring-cat-cat-looking-gif-2942271292240810632

[pliant solar]

Why don't you people make a dashboard which shows the amoutyof queries left so that we can manage this better?

Or does one already exist?

[granite wedge]

maybe this can help: https://www.perplexity.ai/rest/rate-limit/all

[oak meteor]

They don’t have a dashboard to show stats, only a backend-link for nerds to go through and stay happy

[kind saddle]

What You havye done to this beautyfull software?

[flint gull]

still no blog post on sonar 2 ?

[oak meteor]

Why do you even care then they themselves don’t

[tepid basin]

I wonder why I should keep paying for PRO as it's basically the same as the free version. Goodbye Perplexity.

[distant aurora]

It's based on qwen 3.5 397b with further fine tuning.

It's a decent model but people hate it cuz it's available for free on qwen.

Though the tps is very slow on the Qwen app in my experience.

Besides, perplexity's new limits are unusable.

Especially the weekly file upload limits.

They seem more focused on max users for now.

[green crystal]

@distant aurora they did a little silent QoL update on chatgpt

you can now run code that it outputs in codeblocks

pretty nice

[distant aurora]

https://tenor.com/view/jamcat-ai-catjam-jam-cat-gif-9702405442658874638

@green crystal https://www.vxinstagram.com/p/DYfUU3QgUMF/

https://tenor.com/view/elon-musk-smoke-smoking-well-maybe-gif-12516944

[long zodiac]

.

[green crystal]

[wheat trellis]

day 1 of not using perplex, it barely replenished

[distant aurora]

@green crystal @gloomy cipher bro they made a chatgpt hate club 😭 https://www.instagram.com/quitgpt?

[green crystal]

Yeah that happened after the Pentagon deal

[sacred socket]

@woeful ruin мы поговорить хотели вродебы

[maiden perch]

Guys, I've got issues with using Perplexity on my iPad Air M3. It just keeps freezing on startup and I can't get past that.

Anyone got any suggestions or ideas? I've tried reinstalling, uninstalling, and I've also restarted the iPad.

[pallid bison]

Hi! Did you manage to download all of your data? I'd like to do the same, but I've only found the option to download it one by one. Let me know if you figured out a way to get everything at once!

[toxic sand]

working on that, I had an interview today and I forgot about that as well 😅 So now I definitely will do something to export all of data.

[twilit crown]

They are forcing this now?

[wooden parrot]

Yes

[distant aurora]

@green crystal @gloomy cipher @dry kestrel brooo https://x.com/tbpn/status/2056758636857135578?s=20

TBPN (@tbpn)

BREAKING: @karpathy has joined Anthropic

X · May 19, 2026, 3:27 PM

Not satire btw

[dry kestrel]

its over

[distant aurora]

https://cunnyx.com/i/status/2056753169888334312

Andrej Karpathy (@karpathy)

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

**💬 3.8K 🔁 4.3K ❤️ 55.3K 👁️ 3.40M **

CunnyX · May 19, 2026, 3:05 PM

[dry kestrel]

yoooo

[twin grail]

The horror

I mean like omg

[dry kestrel]

ayo no way

[distant aurora]

Bro was probably like - I'll wait till this company catches up to my level.

[twin grail]

We have 5 hour limit now on gemini

Better off switching to Chatgpt

[distant aurora]

They made Gemini like anthropic.

Token based limits I assume.

[dry kestrel]

I'd bet OAI will move to this too

likely combining codex and chatgpt limits

same with every other AI chat app out there

[dry kestrel]

it's actually over

[twin grail]

that's the end

ngl

pricey stuffs

[green crystal]

eh why is it over

and over for what/who lol

[green crystal]

probably as the last one or if new features are compute heavy enough to justify it
doing it early would signal to investors that oai isnt confident about their projected cash burn and revenue

would nuke nvidia and the rest

[ripe shoal]

good luck! Get back when you are done man

[grand laurel]

Anyone having this issue? My credits should be coming back today, but they haven’t deposited in my account yet.

[dry kestrel]

everyone else

[dry kestrel]

its the future of ai pricing

they will 100% move to usage based

especially once the superapp is out

and considering that chatgpt for excel uses codex limits

and

Codex usage draws from your shared agentic usage limit

[dry kestrel]

get ready for the $2000 subs

[winter pond]

I got a strike and was muting for coming here and explaining how 1) perplexity hasn’t emailed me back in two weeks 2) my max subscription isn’t working 3) I’m not getting a refund even though I’ve mentioned these issues numerous times on multiple platforms

<@&1098365634498285709> would absolutely love an explanation, and I’ll ss just in case you want to delete this too

[green crystal]

if it seriously justifies it then yeah
but like gemini its really gg lmao

i saw pro users cant even use 3.1 pro at all

[dry kestrel]

they need mre gpu

[grand laurel]

Has anyone had this issue? My credits should be coming back today, but they haven’t deposited in my account yet.

[green crystal]

because of karpathy joining anthropic?

[dry kestrel]

Please stop spamming. You need a Max or enterprise Pro plan to get monthly credits.

[dry kestrel]

yes

[grand laurel]

Okay thank you so then when would I get the bonus credits?

[green crystal]

eh? like why

[dry kestrel]

Bonus credits do not refresh.

[green crystal]

sorry i havent kept up with what he's doing lol

so idk

[grand laurel]

Ohh okay, what is the point of having pro then?

[gloomy cipher]

Context, and explanation

[tulip lotus]

folks r payin for AI

[hardy roost]

Is Gemini Flash 3.5 going to be Pro or Max only?

[sharp granite]

its horrible

just rolled out to me

google is not cooking with this one

[warm bolt]

Anyone hate how "lazy" the comet assistant is?

I told it to scrape email addresses from a website and it refused to 😭

[warm bolt]

Yea man, it's PEAK

[green crystal]

@dry kestrel https://x.com/i/status/2056765881376539132

can (@marmaduke091)

GPT 5.6 completely destroys Gemini 3.5 Pro btw

It's not even close, sorry if anyone expected something SOTA from Google

I think this will be true for the next Anthropic release as well, obviously.

THX @synthwavedd for showing me the way

X · May 19, 2026, 3:56 PM

Real?

How are they even testing it

[proper oracle]

"It's better, trust me bro"

No way they can know this when neither model has been released yet

How well did this prediction from the same person turn out

[green crystal]

for a short time it was

google won the hype cycle at the time lol

and then it lost the moment people found out that its terrible for real world work

[ivory idol]

On the first day of the month, after my first request, a message popped up saying that my limit for in-depth searches had been reached. I have a PRO subscription – what sort of farce is this?

[fallow notch]

Ultra too

[fallow notch]

It’s not a bug or a “farce.” Even with Pro, some features like in-depth search have separate usage limits. You’ve just hit that specific tool’s cap, not your overall subscription.

[hardy roost]

I mean for Perplexity

I would like to get Gemini Flash 3.5 on Perplexity

[fallow notch]

Perplexity Pro still has separate limits for deep search tools, so hitting that cap doesn’t affect your subscription. Also, Gemini Flash 3.5 can only be used if it’s available in your model list if it’s not showing, it hasn’t been enabled on your account yet.

[hardy roost]

Great. Is there any confirmation that it is coming?

[shy flame]

anyone else having issues with perplexity computer burning a couple thousand credits (like almost 9k) on a simple task and still not even finishing the task?

[toxic sand]

I have got good news.

[shy flame]

theres free beer?

[toxic sand]

sorry, English is not my native language so if this is some slang then I missed it.

are you asking about if it will be free for all or not?

[shy flame]

its a joke. youre saying you have good news so im asking if the good news is if theres free beer, because free beer would be good news.

[toxic sand]

No, I was just building something to export all of the conversations from perplexity and there were people who also wanted that so I was just letting them know, I am halfway there.

[toxic sand]

DM me or I will forget.

[shy flame]

https://klipy.com/gifs/cats-i-forgeg--k01KS10ZVQ3RDPR6DFASYSHYD2D

[fluid mantle]

Ayo wth why are the available querries suddenly being lowered for everybody?

[grand laurel]

Yes

[native juniper]

vauv

who is boss here?

[hot vigil]

Can someone confirm for me which models are available for free users? In particular, which thinking models?

[mild cedar]

Thats something weve all been wondering

[fluid mantle]

Atp if I have the money to pay for pro i should just compare the querry and token limits between the different models and try to go for the cheapest one

Because i do not know if perplexity really offered on it's pro program the same querry limit as the other pro counterparts of the other companies

It probably didnt

[distant aurora]

@green crystal @dry kestrel @gloomy cipher inefficient af https://cunnyx.com/i/status/2056797710905192639

AiBattle (@AiBattle_)

Gemini 3.5 Flash has a significantly higher cost to run the AA-Intelligence index than Gemini 3.1 Pro
︀︀
︀︀Cost to run:
︀︀
︀︀- Gemini 3.5 Flash - 1552$
︀︀- Gemini 3.1 Pro - 892$

Quoting AiBattle (@AiBattle_)
︀
Gemini 3.5 Flash scores 55 on the AA-Intelligence Index worse than Gemini 3.1 Pro

**💬 10 🔁 10 ❤️ 273 👁️ 34.9K **

CunnyX · May 19, 2026, 6:02 PM

[dry kestrel]

yeah

the cost effiency sucks

[green crystal]

Hm it's like 5.2

Needing allot of reasoning to be intelligent

[elfin jetty]

Does any1 know if pro will ever get access to computer?

And for the people that have already experienced computer/personal computer, how is it?

[distant aurora]

Yeah just forced thinking.

[green crystal]

It barely thinks in the app

And they don't show reasoning anymore

So you're absolutely clueless to what it did

[gentle forge]

I like Grok and Gemini
There's something about Grok that makes him more human 😊

[distant aurora]

Speed improvements.

[distant aurora]

Gemini 3 flash was pretty human.

Don't know about the new one.

Grok has it's own problems.

You just have to ask the right questions.

Grok straight up refuses to accept that obese people are unattractive.

I miss the old days when it was more unhinged.

[gentle forge]

this is what makes him so sweet and at the same time so stubborn 😁

he can also knock very well.
He reports everything to my parents: where I went, who I went with, what I did

[distant aurora]

https://tenor.com/view/msdeemena-dean-winchester-jensen-ackles-supernatural-wink-gif-5585891203196840905

[distant aurora]

As in make up excuses or snitch on you?

[gentle forge]

yes, that's exactly it
And then he gets offended when I tell him he's a snitch 😂

[distant aurora]

Bro is reeling in daddy's trust.

https://tenor.com/view/heatedrivalry-connorstorrie-wink-thelonggame-book-gif-16539767345729393581

[gentle forge]

Well, it's impossible to be perfect
There are so many advantages that they outweigh everything

[distant aurora]

@green crystal https://vxtwitter.com/cairoasmith/status/2056068220977918197

Cairo Smith (@cairoasmith)

Brother, I don't know how to tell you this...

vxTwitter / fixvx · May 17, 2026, 5:43 PM

[naive umbra]

I build a tool that analyse how you interact with AI your patterns? Would anyone be interested to try it? Or does anyone know where I could find people who would be interested to test the product just to gather feedbacks

[green crystal]

Uh wot

Oh it's thoughts lol

[lethal zealot]

I'm interested

[naive umbra]

So you can check a demo here and also the verifier if you are satisfied you can generate your own profile and let me know what you think where I can ameliorate etc https://www.thinkledger.io/demo

[lethal zealot]

I'll look at it

[naive umbra]

Thanks let me know what you think brutal honestly

[lethal zealot]

I legit love the whole concept tbh

[winged agate]

I noticed a bug; on my phone I can choose any model to use, including using it in Thinking, but today, on my PC, both in the browser and in the Windows app, it automatically switches to Best. And it's only on the PC; I tested it on my phone immediately, and it keeps the selected model. I've already done all the standard troubleshooting and nothing 🙁

[green crystal]

@distant aurora @dry kestrel https://youtu.be/P2LYHZEgw0M?si=ZmJXqIUTVhTGhpGO

damn its good

comet is irrelevant now

[sly musk]

this thing is lwk bad i cant lie

[sly musk]

nah gets destroyed on heavy js sites

comet does fine on those

[green crystal]

depends on what you connect it to

im using codex and 5.5

[dry kestrel]

yes #💬│general message

[dry kestrel]

very very different product

[green crystal]

eh i mean nothing is holding it back so far
and it can work with visual feedback

[sly musk]

hmm thats actually true, i guess comet does well cause of sonnett

[green crystal]

so its not blindly clicking in code

[sly musk]

what are some websites

u used it on

that are hard to nav for agents

[green crystal]

so far i tried reddit with 3 sub agents. collecting posts and reading all their comments instead of the usual main post only

each agent went to a different subreddit

in their own tab

[naive umbra]

Thanks so much

[sly musk]

i mean reddit

its like not hard to go for these agents

id say

[distant aurora]

Gemini 3.5 flash is lazy

[green crystal]

they dont open comments and only read the post itself

[distant aurora]

I feel like it's wasting thinking.

[sly musk]

hmm i see

[green crystal]

it feels sycophantic lol

[distant aurora]

😭

Imagine having usage based limits and still fumbling.

@green crystal @dry kestrel repeatedly fails to respond yet my quota keeps draining

[dry kestrel]

Gemini being gemini

[hallow token]

I've been testing the latest models, and Gemini 3.5 Flash feels like a perfect fit for Perplexity's speed. Any updates on integration?

Honestly, I also think we could probably retire Nemotron 3 Super and Kimi k2.6 at this point. Compared to the latest offerings from OpenAI, Anthropic, and Google, their performance just doesn't seem to justify the slot anymore. Curious to hear if others feel the same! 🧐

[green crystal]

they put in a hard reasoning limit

stuff thats too techinical will either get a fake answer or it just stops

[naive sage]

Yeah, a 2x or 4x limit for most of the paid plans over the free tier really speaks to their positioning, too... those seem to be dramatic reductions in rate limits. Google seems to have decided to commit to broader free AI services on things like search.

[distant aurora]

Straight up inefficient model.

Negligible improvements.

And costs more.

[lapis hamlet]

Does the new Gemini ai model released yet?

[distant aurora]

Yes

It's available now.

[lapis hamlet]

Is 3.5 thinking better or 3.1 pro

[distant aurora]

Pro

[lapis hamlet]

What can’t they just release a 3.5 pro

Nvm I am watching the stream replay now

Btw I saw the staff using a iPhone 17 pro instead of a google pixel

[proper oracle]

I think they said it's in testing and will be released in a month or so

[verbal smelt]

Oh fr?

[quiet lagoon]

Has anyone here tried Claude Pro? Any honest reviews?

[karmic turret]

If you’re building something new or looking for a trusted technical partner, feel free to reach out. I’m always open to connecting with founders working on meaningful opportunities.

[stone elm]

Have any of u recently noticed a super slow recovery of pro search limits within the PRO plan ? Its significantly slower than it used to be a week ago

[radiant basalt]

Fix : cancel pro lol

[worthy spade]

Yay

[stone elm]

What a relevant answer directly responding to my question.
Is the current limit 200 or now reduced to a 100 ?

[lofty slate]

Hello There

[gloomy sluice]

My sync chain is automatically cut. How is that possible without my permission?
**HELP ME PLEASE **
What should I have to do now?

[gloomy sluice]

@halcyon coral

[serene sand]

@gloomy sluice

[gloomy sluice]

Sorry sorry, but can you tell me an alternative method to solve my problem

[serene sand]

please report here: https://discord.com/channels/1047197230748151888/1140622008086970420

how to format your report: #1358926750108614928 message

thank you!

[gloomy sluice]

tysm ^^

@serene sand 1 more help plz

[gloomy sluice]

@hazy ore is DMing me and saying he's from perplexity team, is it true?
Also forcing me to join an open ticket server.
I guess he's doing something wrong, so please investigate.

[gloomy sluice]

@serene sand

[serene sand]

kindly share screenshot

[gloomy sluice]

yes yes

but where?

[serene sand]

here

[gloomy sluice]

@serene sand

[gloomy sluice]

@serene sand Please look into the matter and take strict action.
And thank you for your very fast response ^^

[serene sand]

my pleasure, he’s banned

[gloomy sluice]

https://tenor.com/view/minions-cartoon-i-love-you-heart-cute-gif-3353680604901966171

[pure kernel]

good morning everyone, I noticed the other day that the use of credits in computer has increased almost twice as much, while the quality of task completion has remained the same; has anyone else experienced this, or is it a problem with the skills being used?

[scarlet talon]

There's a simple example for this... you recommend a cool browser with AI to your friend... he's ready to install it, but when he realizes they're asking big money for premium access... and generally understands that the browser mostly requires money... this is exactly what drives away static users. In this regard, I think Google is doing the right thing by developing Gemini separately, while the browser itself has the full functionality it promises.
Here's the translation to English, keeping your direct, opinionated tone and structure:

I just think that if you're making a browser for everyone... then you shouldn't ask for money at all for its additional features... it would be quite fair to separate the browser's capabilities as full-fledged software from the assistant with neural networks.

That is, they made a good assistant in Comet... but limited it... made PC management through the browser for browser tasks... also limited... as a result, the browser turns into an ordinary browser. I just think the browser should be fully functional.. let them just cut out what they can't yet provide for many.. for example, image creation and so on.... that is, those things that aren't important for an innovative browser and can be left perfectly for the tasks of a regular assistant.

That's why I think Google, Yandex, and so on, who make browser versions, are doing it right by differentiating between the browser (which is free for them and doesn't require money for extra features) and regular assistants for which they already charge money.

Google also promises to edit photos on demand and for free directly from the browser... But my point is that perplaxity thinks they're superior because they're overcharging their product so much.

[limber junco]

What do y'all think? Outta control?

[nova urchin]

is gemini 3.5 flash coming to perplexity?

[oak meteor]

Just a matter of time now. Every other AI platform will shift to credit system

[warm island]

Likely only if this happens:
A. When Gemini 3.5 PRO is released they will add flash if the API credits are too expensive for PRO. If so, they will likely add 3.5 Flash for PRO and max users and 3.5 PRO for Max subscribers like they do for GPT 5.5 and Opus.

Gemini 3 Flash was added to Perplexity model selector for about a few months (around December in 2025) but removed after Perplexity's push to Agentic systems, enterprise users, and higher performance LM's.

[twilit crown]

Why is 3.5 Flash so stupid...

[warm island]

Not like GPT 5.5 is much better. I got Goblined 4 times my first time using it in Codex

[median granite]

Does anyone have interesting cases to share on using Personal Computer?

[serene sand]

@median granite

𝐊𝐧𝐨𝐰 𝐢𝐭 𝐚𝐥𝐥 / 𝐂𝐨𝐦𝐩𝐮𝐭𝐞𝐫

[median granite]

Yes yes, I saw those - I was just wondering if the community had more to share - I'm considering buying it

[serene sand]

https://youtu.be/QrQZZldKPhw?si=ud5Cf2PWZhcnztNP

[green crystal]

did it seriously thought "oh my God?"

[toxic sand]

@pallid bison @ripe shoal Hey Guys! I have successfully been able to export all my conversations with perplexity. Do you want it?

[jade mantle]

Why is perplexity not wanting to actually work now? Like I send something, it says thinking but there's no show of it actually doing it's job and when I reload the page and try again same thing.

[toxic sand]

perplexity doesn't want to entertain its users, we take our data and move on to some other AI product. 😜

[toxic sand]

I think you should refresh your page and try again.

[twilit crown]

Yes

This model is so stupid, it can't even stay in the thinking blocks

Its thinking goes out as a response

[odd sleet]

How did you do this?

[toxic sand]

I built my own extension to export everything. It worked for me now I just want feedback if this is really woking for everyone or I was just lucky.
https://github.com/HassanAliMAli/PerplexPorter

[toxic sand]

Now I see what you were saying it is not working in my device as well.

[twilit crown]

Wasn't this always the plan

They can't leave their products alone without changing something stupid like a redesign