#💬│general
1 messages · Page 170 of 1
Same
I just switched to chatgpt lol
o3 with a more fleshed out prompt beats current ppxl deep research
@distant aurora tool usage during reasoning is such a nice difference from ppxl
Even o4 mini does it pretty well
Yep.
Og is og.
theres no claude opus?
I need a hug
Hey yall, I'm a bit lost with the new update to the app on Android. The usual voice options have all been replaced, and the app no longer allows me to select or change the AI model. The setting now just says "Select model during query or follow up input." Huh...??
I am receiving messages to sign up for the Pro version event tho I do have the Pro version. I am also being asked to log in multiple times each session on my iPad Pro. Please advise.
Model picker was moved here. I don’t have an android but it should be the same process: Press “ask anything”, “search” and click the “best” (or whatever model it says)
Thanks matsku for your reply. I only ever use the 'hands free' option so didn't notice there was an option here when manually inputting text.
If I select a certain AI model in this fashion, will my choice carry over when using Perplexity hands free, which currently no longer has its own option to change the AI model as it used to?
Probably? I haven't used it personally that much, so I can't say, but I would imagine you can try it by choosing a reasoning model and see if it takes longer than a non-reasoning model
Or just Claude 4 vs Sonar too
Sonnet Thinking has been nice in PPLX
Why do I feel like Claude 4 sonnet gives a lot more coding than other models? Like it just keeps on going and write a really long code unlike other models
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
They sacrificed everything for coding performance.
Lol.
i made myself a new "new tab page"
Lovely
Seems like everyone's using brave nowadays

little bit explantion pls
oooh
Hello world!
man i wish theyd fix my widgets
defaults to this shit solely cuz its a student acc
useless
is thst in browser or app?
Am I supposed to report this?
absolutely
yo what is this deeper research
how do u have this yellow theme?
complexity extension
ooh lol
yeah i saw this just now too while hovering over models
It was like this for me too when I deleted my cookies, don't think it's just a student thing
this didnt happen on my non student acc though
its been like this for the entire duration of me using this acc
New feature..?
i dont think so, like it used to show the weather and a few news articles
i liked that
wtf is this lmao
and these buttons arent even "features" they just add prefixes to the prompt bar
its great for people new to the app but they should phase it out and give u the normal ones later
deeper research really gave gud response
but it was not long
gave just 20 questions and answers
"deeper" my ass
How deep is your ahh?
Average conversation with grok
That problem was interesting, so many top tiers were having issues with that one
Qwen lowkey overthinks it.
So I don't think it can possibly go wrong.
They probably put self-assessment and self-criticism in the system instructions.
qwen does that naturally
qwen just like me
tampermonkey extension allows scripts, so i made a script that adds quick links and my calendar agenda to perplexity.ai and set it up as my homepage in brave. Now, everytime you open a new tab, you see this
wednesday
^
Happy tuesday
Wednesday or Sunday
Yeah could be Sunday too.

Anyone else here who has comet? 👋
same question
nope
i am in waitlist but they are not sending me
invite
I’ve used it for the past week and it’s amazing
It’s replaced arc as my main browser
ah i need invite\
😦
can you show me how it looks
When are we getting Claude 4?
already there
👋 @vague pagoda 
Heya!
Did not realise there was a discord!
tell me some use cases of ur exp where comet was very useful
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
google has alr launched veo 3 and they give only 5 video limit of veo 2 for advanced users
pathetic
noice to mee you 🙂
hello
I don’t see it on iOS
Hello, good people! I am looking for a dataset that can be used to train a model and alter its identity.
For example, if I use Qwen2 as the base model and fine-tune it with this dataset, I want the model to respond with a custom identity when asked questions like:
"Who are you?"
"Tell me something about yourself."
Ideally, the dataset should include prompt-response pairs that enforce a consistent identity (e.g., name, backstory, or capabilities).
Disclaimer:
"This video includes content from Rick and Morty, a show owned by @adultswim and @WarnerBros . I do not claim ownership of the footage or audio used in this video. All rights belong to their respective owners. This video is made for an opinion, thought, or reaction and or comment, and is shared under the principles of fair use."
...
what is best solution for vibe coding in 2025 may?
There is no such thing as the best one, but you can do your research and find something that suits what you are looking for.
cursor, cline suck
You have to try it.
Just watched 7 seasons of Suits this week haha
In a week?? Insane
The grind is real
When will the perplexity app for Mac get better
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
Pls add this for android #📢│announcements message
@vague pagoda Here the link is not correct
nice find! thanks
just wondering how do i use it? https://www.perplexity.ai/properks
takes me to this page
should be an option in there
Note that Pro Perks are available exclusively to Perplexity Pro subscribers with a US-based subscription
Refer to https://www.perplexity.ai/help-center/en/articles/11321927-perplexity-pro-perks for more info
oh sed, i am from canada
And what is this new button?
Reasoning likely
Deeper Reasoning?
Projects
why doesn't it appear in my perplexity homepage? is it a beta testing thing?
Temporarily unavailable as we work on some improvements to the redemption experience. We expect that this Pro Perk will be available again later this week!
I think so, as well as tasks
alright thanks!
fixed ✅
Please look at the DM
i cant see the new button for some reason
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
It this was regarding that assistant whose name starts with an "A" and ends with an "a", we would need separate storage just for those wrongest answers. 😂🤯😶😶🌫️🫥
what is the model with longest context window? I want to translate something
I just tried Grok 3 it's got the ability to communicate in multiple languages on its own. It doesn't interrupt as much either. You can also type quite a bit. It's pretty impressively done. Perplexity is still smarter, but this is processing. Next 5 years are going to be wild
i think regardless of the model, perplexity puts a 32k token limit on all models. correct me if i am wrong
you can use gemini 2.5 pro or GPT 4.1 outside of perplexity for that.
To be honest I dont notice much difference with reasoning gone
I didn't really read it properly anyway
But yeah its OpenAI that started this bs
@green crystal https://www.rxddit.com/r/perplexity_ai/s/66w1xhMyQY
New sonnet 4 is soo much worse, it make far shorter answers, ultra concise, no imagination or creativity, don't take any risk, ignore tons of stuff in the instructions an,d will often refuse to answer for absolutely nothing
So could we get back 3.7 please ? anthropic still offer the API and it's the same price
yo guys can you redeem the Cursor pro perk? it only redirects me to the settings page. All other Perks are working just Cursor is not. https://www.perplexity.ai/properks
Not currently available > #💬│general message
Wait. Fr?
thank you!
@tame current
Well openai is the one that decided they should conceal their reasoning with summaries
Like there was literally no point
Claude showing it all mfmfmggg
Claude showing all the ways it goes wrong.
How do I use Perplexity Projects? I saw it as an option and then it disappeared?
I do like this Finance Dashboard on PPLX. Is there any way to get it to update faster?
Anyone know where to access my pro perks
Are you on mobile or web?
Web
If you click your profile icon in bottom left, it will be in that menu or in the "View All" section
I see it but it directs me to my account details page
The Pro Perks link directs you to My Account?
Yeah exactly
And you are sure you have Pro?
I don't think so. That is weird. I would submit a #1140622008086970420
I'll try on my phone then if not Ill do that
Good luck!
What about on phone where do I find the pro perks
iOS or Android?
iOS
Might not show up if you're not in the US, as they are available there only
Ah ok
Somehow the Claude 4 Sonnet didn't impress me
Why aren't more people talking about this
Wonder if they are beta testing it
How Projects differ from Spaces?
I tried running with it but somehow it instead went to research
Both are unrelated
This is under search and research sections
This is apparently for building charts, apps etc
Should've taken a screenshot when I had the chance
People argue about models hahaha, it's funny.
Well, I've been testing it in reasoning mode and it runs better than version 3.7, it interprets my prompt better and is more detailed with good narrative.
Any one knows context window of Claude 4 in pplx?
32k
Really?????
Or deep research.
Some sonar model had 200k I think.
DR uses their own "models" (probably) so they can afford to put it higher
But no one uses sonar so.
Didn't they say 2.5 Pro was 200k?
R1 1776 32b? 💀
Damn 32k is almost nithing
If working with documents, yes.
Its how they can afford all these models
Grok has 130k. 1m only works in deepsearch.
A useless model
Well they are likely majorly finetuned since then
Sonar models now
Which is just like "Everyone else's perplexity responses from google models distilled onto llama" or something, dont know
My point was that it's as dumb as a 32b model.
Whatever deep research uses.
what happened to the Perplexity Browser?
I thought it's available to mac users now.
It’s getting close to the finish line
I need to update chatgpt about the status of my cough.
Maybe I'll export the chat and upload it to ai studio.
Ok I will be mostly ok with opus 50
Opus 4 50
Give me opus or give me death
Oh sweet child, long ago there was a time when unlimited opus was promised to the users of perplexity, then overnight it was taken and replaced with a mere 50 uses per day, that day will forever be remembered by those it affected
Ahh, yes, perfect, thanks 🙏
Opus doesn’t sound like a real word anymore to me
Opus ❌
Copus ☑️
It’s so much more than a word now
Is it a lifestyle
It’s the meaning of life
Rich ppl lifestyle.
Opus 4 should be a human right
Lol
Perplexity on it's way to make a 100 dollar subscription after this post
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
We shall not be silenced by the offer for free hoodies
And you are not doing well with Sonnet 4 of reasoning?
It’s the principle, give me opus or give me death
They've said about 03 and they haven't put it yet...
Fuck opus
Make quality deep research a human right.
Lol.
Idc what provider it be.
Watch you mouth, opus 5 will be my ai wife
Exaggerated
Then some random dude will jailbreak your wife.
Nooooooooo
That's how you lose half of your property.
😭Lies all lies
Calm down man
🅾️🅿️⛎💲5️⃣0️⃣
why shouldn't I sub to t3 chat when they get all the latest models instantly?
pplx is yet to get opus and o3
How much usage do you get of opus?
its not even about opus as much as it is about o3
Ok np, yeah t3 looks solid, I think that perplexity research and deep research is pretty solid, I use that for random questions, if you don’t use that as much t3 might be as good or better with its being so cheap
yeah
I use openwebui for my normal chats now days
hard to stick with perplexity ig for me
I still have it free so I will figure out what I want to do there later, haha
If you check out t3 let me know your experience, also @sleek vortex is working on uncovr ai so hopefully thats ready by time my free runs out
yeah sure
what's uncovr ai btw?
https://fxtwitter.com/vishyfishy2/status/1926051179068330434 it’s a ai website he is building, you can see his progress on X
another ai chat interface
smh
there are sooooo manyyyy ai chat interfaces now a days
Yeah but they all kinda suck, that’s why I moved most of my chats to my own api use, his so far is pretty good
why do they suck?
Joe Biden had lobotomy
Taylor Swift has her save environment bs while travelling in planes.
I don't have to explain ice spice.
One example is You.com, it has annoying file upload over 5k characters making it unusable for me, they removed the control over full context vs snipers in web search. Poe is ok but not great imo. Perplexity use to be really bad with file uploads but not it’s mostly fixed but the context window is not great. Your takeaway can be that I’m just very picky lol. But it’s hard for me to find a chat app that has all of the models I want without restrictions on context window, or bad UI
Lmao. My kind of guy.
ChatGPT, Claude pro, and Gemini are solid, but bad model selection for obvious reasons lol
Super grok is overpriced

I thought you.com had 64k context limit. Can't relate to the 5k characters part. I uploaded an entire WhatsApp chat export.
No I mean if you paste in more than 5k characters it will upload that as a file, not a issue for everyone, but bad for my workflow
I had you.com temas so I had full 200k, which is undeniably a good value, but you know I’m picky haha
Always be picky.
Which ones don't have this issue btw? Asking for knowledge.
yeah thats horrible
number 1 thing i hate about you.com
but ig its kinda how they can afford to let you use 200k context window
I don’t actually have a go to recommendation, they all have cons
native gemini or chatgpt sub
and i think claude as well
I'm asking specifically about the 5k input limit.
Oh yeah, on perplexity it’s easy to override the file upload
I doubt this.
What are the pro perks where can I see this in the mobile app
I don’t use chatGPT or Claude anyone but I think they allow more that 5k characters
I google ai studios is goated with 2M
I vouch for grok and Google.
You just run out faster if you use it, but yeah you get full 200k
and i think flowith has some pretty good limits as well
Woh, did now know 500k was a thing they did
yeah the usage caps make the $20 unusable
was running out after like 20-30 messages i thik
I have privacy concerns regarding how and why it found my qwen chat specifically.
wdym
Do you guys just go with the default pro model or which model do you use - idk which ones to use for good results hence the default one works just good enough.. Thoughts?
o4 mini or claude
can you share the chat?
And the default one does it pick Claude or their own model? Any idea
it picks sonar which is shit
I think it uses sonar (llama)
And Claude or o4 are way better than sonar is it?
I like sonnet for the default model, but o4 mini is also a great option
yes it is
They’ll be under your account page, don’t think they are on mobile though, + they are US only
Claude and o4 can search web and give references or nah?
they can
I’ve been running Claude / Claude reasoning lately, not a too big fan of o4
tbh i havent been using pplx lately, mainly just for comparison purposes
and not for actual work
o4 is my go to for math and stem, but everything else for me I would say sonnet
For me it's O3, grok, Gemini.
youve shared our qwen chat before so its accessible via the internet and due to your prompt being extremely unique, the provider found it...
I've been using default sonar for like the past 4/5 months lmao need to start using sonnet more
Well yeah but o3 is outside of perplexity, but yeah it is better at math
But I pasted the same thing on perplexity and chatgpt and Deepseek too. Why dig up only qwen?
I pasted the same thing on grok too.
In chatgpt theres an option for the link to appear in google searches
and I think its just the API provider that they are using
which manages the searches
I doubt that they built their own search engine
Also, I ran the same prompt twice, in first attempt, it didn't even look for qwen.
Lemme share that.
Are all the latest Google models on pplx btw?
Yeah LLM results wont be stable.
it depends on the search queries that the LLM decides to do
Just Gemini 2.5 pro, flash and flash thinking are not
its like when you run deep research two times in a row with the samae prompt
and they get diff amount of sources
And they have no Gemma models, but I don’t think you want Gemma anyways
Also, even then, the result was incorrect.
What is the answer then
And qwen chat ain't even a reliable source.
What models/providers have gotten the right answer then
Grok, perplexity, Deepseek.
Chatgpt
Yes.
i mean if you can cough up the $$ for api pricing, i can make the interface do anything really
lots of them are bad just because of UX (e.g you.com, somethings on pplx)
Did you enable reasoning?
Yeah. Your bad. https://chatgpt.com/share/68310d78-b750-8013-a252-d538c2d0064a
Drh and oden shouldn't even be there, lol.
Xm4 and type 19 make some sense.
I'm assuming the non reasoning ignored damage drop with distance.
Wait
Muahahaha
t3 also doesn't give you opus or o3. it's just bring your own key
what do you expect for $8 though
pplx loses millions of $$ giving out model access at practically unlimited rates
I wasn't expecting anything, I was informing someone who misinterpreted the marketing
but what I do expect is more transparency, as you can see how most people think that it's more capable than it actually is
none, its bring your own key
ppxl pre prompt is probably screwing it over
Hey does anyone know why the pro perks are not working for me?
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
Im pressing on the pro perks button and its just bringing me to the account section
And also shopping section is not working at all i press on the shopping button and it doesnt do anything
same here
Hey, why did the Cursor Pro I claimed on Perplexity just disappear out of nowhere?
Are you guys in the US? @severe bramble
I got it to work now but i dont see all the promised perks
Which ones were promised
oh you are talking about: https://www.perplexity.ai/properks
Yeah
idk lol haha.
No cursor no perplexity supply 25$
They might have had to temporarily remove them, hopefully will be back
temporarily unavailable
Okk thx
Hey, bro! The Cursor Pro I got from Perplexity the day before yesterday was gone when I opened it today.
Lol please read two messages above this one
asia
It may not be available in your country. Not sure
"Redemption experience" 🤔 My hunch is Cursor wants to renogiate the deal as even accounts with redeemed perk have been stripped of Cursor Pro benefits
guys see what i got
How did you obtain it
O3 is still having trouble with it too
That’s why o1 pro mode and deep research exist 
connect to usa vpn
then go to https://one.google.com/join/ai-student
is need more help then you can dm me
@distant aurora
I am no longer a student
haha xd
The memory feature where
May I ask if there is still a usage limit for the pro version now? Using Gemini 2.5 or Claude 4.
In perplexity I believe you have 600 uses.
Can I just say I'd rather the AI help me sort through discounts I already have available
600 uses per day of perplexity pro? or is that for other algorithms
600 ?
If it's 600 times, it's completely sufficient. Generally, frequent users might use it around three to four hundred times a day. For high-frequency use, like searching or writing articles, the daily usage could exceed 400+.
Per Day
Right, ok
Claude 4 crazy
Ud have to be hitting it all day..
Is Gemini 2.5 Pro 0506?
Even for me I barley pass liek 150 in 6 hours
Haha, I've been writing a short story recently and have been heavily using Gemini 2.5 Pro and Claude.
Claude for a story is crazy
Actually, it's mainly used within a certain period, probably no more than three or four hours of usage, and the number of times should be around 200. Five or six hundred times probably won't be used up.
Gemini 2.5 Pro is also very good at writing stories.
I recently claimed the 3-month Cursor Pro promotion offered by Perplexity, but today I found that my redeemed benefit has suddenly expired and I am unable to use it anymore.
Could you please explain the reason for this? Additionally, is it possible to restore the access to the benefit I have already redeemed?
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
Perplexity new feature Project
Is there a beta app?
I personally from what I have read online is the offer was being ?abused? somehow as there was forums somehow reselling this and its possible either they were exploiting something to get the offer discount and then selling it and it ended up becoming a huge issue as they removed it fully and would be a good reason for them to fully remove the promotional offers from users using it (Even if it wasn't 3rd party resale and legit pro members like myself) since they say it should be available again later on this week?
Actually, I bought Perplexity mainly for Cursor, but now Cursor has been revoked.
very pleased with the new finance page
@distant aurora 👀
For anyone wondering about Cursor:
Unfortunately, we've seen a high amount of abuse targeted at the Cursor discount offered through the Perplexity Pro program. At this time we have cancelled any existing redemptions of the promotion, and accounts have been reset to free trial status.
Best regards,
The Cursor Team
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
No, this is a normal application, this happens sometimes, that is, if you re-enter, then this function will not be there again
There was access
I don't know, each person can be given access to one session at a random moment

I liked it
you dont need student mail i did with my personal mail
I bought a one-year Perplexity membership support three-month cursor pro the day before yesterday, but I was informed that my cursor pro has already expired. Is this malicious fraud?
Why not release a Linux desktop version? And I see there are no plans in this regard. Have your business directions and plans truly considered this?
it never ever worked i tried on first day
i guess you should cancel it and take a refund if you particularly bought it for cursor
It happened again, but the projects don't work, instead of them works Research
got access to flow
can you translate
Search
Research
Project
ah i see
good good
Bro check ur dm..
ah ok
Hi all! Was wondering if I’m the only one who sometimes finds Perplexity’s answers a bit short and less datailed? Especially if compared to chatgpt.
Kinda new to perplexity so might be just my own fault.
try using deep research and different model
Understood, thanks.
Wonderful, thank you!
happy to help you
it will be great if you thank me by vothing here
Yeah that works fine for me even with you.com and perplex
hmm
Done bro.
gemini worked?
ok
Me right now
😂😂😂
This will hit hard on the tokens. Most people won't be able to afford it anyways imo.
Some guy here tried it earlier on grok with deepsearch 😂
It started thinking it might be a date or ph.

Told it to generate an image of our conversation dynamic. This is old.
I was able to successfully redeem the perk & had access to Cursor Pro up until whenever they disabled it yesterday
hmm
i just used my edu mail to get cursor yearly
i was fast as they removed india the day after i redeemeed
我可以用这个干啥
Hi all!
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
perplexity has how many windows of context
hello
How are you?
Hey, is there like a channel for Comet testers?
sadly no
hi guys does anyone know of a good image2image generator that can do fantasy creatures? i need picutres for my dnd bbeg in various environs
@tame current https://www.instagramez.com/p/DKBZ4x0Ju3F
No way it failed twice.
As of May 24, 2025, India is witnessing a gradual rise in COVID-19 cases, prompting health advisories and heightened preparedness in several states, including Delhi, Maharashtra, and Karnataka.
Delhi
Delhi has reported 23 new COVID-19 cases, marking the first significant increase in nearly three years. All patients are stable with mild, flu-...
176017
first one is working ig
No, it should have loaded the images in the chat.
💬 896 🔁 0 💜 175.9K 👀 0
As of May 24, 2025, India is witnessing a gradual rise in COVID-19 cases, prompting health advisories and heightened preparedness in several states, including Delhi, Maharashtra, and Karnataka.
Delhi
Delhi has reported 23 new COVID-19 cases, marking the first significant increase in nearly three years. All patients are stable with mild, flu-like symptoms. In response, the Delhi government has issued an advisory directing all hospitals to ensure preparedness by securing adequate beds, oxygen supplies, essential medicines, and vaccine stocks to manage any potential spike in cases. Hospitals are also required to report daily COVID-19 data and send positive samples for genome sequencing.
Maharashtra
Mumbai has recorded 95 COVID-19 cases so far in May, a significant increase compared to the state's total of 106 since January. At least 16 patients have been hospitalised, and transfers from KEM to Seven Hills Hospital are underway to…
like this ?
Yep
Dhoom 5, lol.
💬 448 🔁 0 💜 25.6K 👀 0
Over 100 Parrots Dead, 50+ Injured as Violent Storm Hits Jhansi, Uttar Pradesh
A severe storm hit Jhansi, Uttar Pradesh, late last night, killing over 100 parrots and injuring more than 50. Heavy rain and strong winds uprooted trees and damaged property across the region.
You live in city
Gemini 2.5pro is like a beautiful village girl who can cook
Claude 4 Opus is like a metro city high standard beautiful girl who earns but cannot cook
I take Gemini 2.5
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
Claude 4 is also that type of girl to instead of learning to cook spends her time on writing ai safety blog posts
And how dangerous she is
me 2
This is how the 2D battle map is displayed in overhead view using Sonnet 4 in reasoning mode in Perplexity. It's much better than the previous, more detailed version.
The text is in Spanish, but you can translate it.
Single man is better...
Hi, pro perks are only for US?
Why Google Flash 2.5 is now without thinking in the Gemini app
Yep
VPN can bypass problem? Because in europe we have no ratio to continue paying for a product that doesnt gives more
good evening guys
Use a VPN and connect to the United States
I tried it the other day and it showed me a code, I didn't use it.
ffs 😦
any update on cflaude? it is clearly not claude
why there is no Cursor 3 month free on Perks?
So you can see that it works using a VPN in TurboTax, however, since I don't need any of these products, well, I don't use it.
Disabled for now due to some issues, they’re working on bringing it back
If any of the mods see this,
You know how you can adjust playback speed of a video on YouTube?
I was wondering if you guys could add that feature to reading the text aloud?
I know the perplexity app has a feature to read responses out loud, but it talks too fast.
I have ADHD and it’s difficult for me to focus, and when I pause the read out loud feature, I can’t make it go back to the beginning and read it again, it stays where I paused it. So I have to exit the question and open it again for it to restart from the beginning. But aside from that minor issue, the playback feature would be extremely helpful for those who don’t process information as fast as others.
So is it possible to add playback speed to the read out loud feature please? 🙏🏼
Will Perplexity have its own keyboard like Gboard or SwitchKey but made in Perplexity?
Should post this in #📱︱feedback-mobile
(or something like #👉│feedback-general / #1111765634267742408 if not only related to the mobile app)
@tame current https://youtube.com/shorts/7GYihKiuQVg
Thank You for Watching this Video, if you enjoyed it Subscribe to the Channel for more.
You can Find me on other social Media Accounts where I constantly Post Memes, Informative Videos and Fun Gaming Content:
📽️ Main Channel • https://www.youtube.com/channel/UCIGHcsO-LizgPtUgmd9-XFw
💰SPONSORSHIP • https://yt.openinapp.co/rd24o
...
Wtf
Twitter is down 🕊️
I’m the #1 user on there
Aaah
7 day shift here we go
@serene sand
If you use nordvpn, it can sometimes give you nordvpn for free for a fat minute
please contact support@perplexity.ai
It’s been a year for me and still no purchase for premium necessary
Wtf
Should i mail them ?
Yes
Ok thanks for the help
@short narwhal https://youtube.com/shorts/YkIGHZp7eDo
#pets #cuteanimal #cat #cutecat #music #puppy #dog #piano #classicalmusic
I have checked it and the Claudes in perplexity put you emoticons without asking for it, however, I checked it in another side and it does not add emoticons. It is clear that it is modified or it is another model.
I use another VPN service, I bought the NordVPN service for nothing.
Nordvpn doesn't work for me.
i use surfshark vpn
It is just bad, i think there is an error, but i dont know why it hasn't been in announcement or something
I don't like that company.
I tested the two models on Claude's perplexity and they do the same thing, however, on another site I checked the writing is much better.
You are missing real servers in needed locations, not virtual ones.
which one do u like ?
I did not receive my hackathon api credits
Can somebody tell the process or help
Hello everyone, are there any ultimate prompts for Perplexity? To make it work perfectly, I need to input a large amount of information and have it analyzed)
Need a wife like this 😭 https://youtube.com/shorts/2jFZkaaPuEI
this sucks.
Yeah bro.
Are there any creators who promote AI software? I want to find someone to cooperate with
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
hi hammy
whats going on
feeling sad
thats the only reason?
umm
and i am jobless
study over?
whats your qualification
skipped placement programs
IND
i regret it everyday
Don't get demotivated. Those programs are worth sh**
apply randomly on linked in
https://f.igerman.cc/u/6fim22.png hey can the search suggestions not be this horrible
theres like one that's not american politics or war
news != search suggestions
you have a separate page for them for a reason
turn off auto suggest in settings
Kindness is clearly overrated these days
I just found a new feature a while ago, but it disappeared under the name 'Project.' The description was 'apps, docs,....' Did anyone see it?
I wish Perplexity was more transparent with the models it offers....
yes it was discussed you got luck that you got a chnace to see
its all based upon luck
it gets removed for you maybe it come back agian
all depends upon luck
surfshark
Paid or free?
paid one
How much
surfshark dont exist for free
There is a place where u can get 2 month surfshark for just 3rs
Chatting on discord
let me guess
you must be chatting with a cool guy
maybe his username is richkeed
he is a smary guy
i know him personally
Who hand washes their clothes here?
Why does perplexity have so much hallucination...
Processors and battery capacities of many phones are wrong.
maybe high?
me when i was in hostel
How much detergent powder should I add to 4.5 litres of water?
it depends on number of clothes not water
tell me how many clothes do you want to wash
and are they bulky
I thought it was based on concentration. For the soaking before the scrubbing and rinsing.
1 scoop is enough
or if your scoop is small
use 2
I'm assuming that would be around 2 spoons.
o3 are they planning to add to the app?
Enjoy
Stability issues resolved?

It is paid
I have a NordVPN account and I rarely use it, so I don't know what to do with it.
@vague pagoda Can a personal account be raffled here?
as in a little giveaway for your Nord account? while that's very generous of you, not really appropriate for us to facilitate it 😅
Yes, it expires in 2026. At least, it's help for the community; if it can't be done, it's no big deal.
Agi is much nearer then we think

Does anyone play Minecraft here?

@tame current hey
WALLPAPER
Might wanna upscale it.
Is there a way to increase text size on iOS?
Lol. Usually people ask to decrease.
You can do so in the iOS accessibility settings
Increases all text everywhere though
I thought they particularly meant perplexity app.
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
Probably, but not a thing :(
I had such a weird dream with my crush in it.
Lol.
I don't even consciously think about her anymore.
Bro, it had a teacher in it who called me baba for my haircut. 😭
Yes, doesn’t seem to increase on the perplexity app text just headings
I use web version cuz of this.
Because of the text size?
Get ready to go back to where IT all began... 🎈
#ITWelcometoDerry is coming this fall to HBO Max.
About Max:
Max is the culture-defining entertainment service for every mood. With a variety of genres that include your favorite series and movies from iconic brands and treasured franchises, it delivers irresistible stories every time. From r...
tell me for mostly what you use perplexity for??? Me Coding
at this moment trying to use for everything... time to learn with Perplexity. To be honest thinking about pro version try for a month
I've been using pro from last year all I can say BESTTTTTT SHHHHH
What model do you preffer for your use case?
sonnet
best
with thinking context is 🔥
more than 2000 lines of code in one go
and for code continuation this prompt works like charm "give me remaining code from where you left of in previous conversation"
Yes.
Web search.
How perplexity is working, is there own database web crawling
Which browser? Using zoom on safari makes the text box massive
Change text size in accessibility in chrome.
I think firefox has it too.
Jack off
nah
How old are you?
Liquor helps sleep.
i dont drink
Vodka to be specific.
nor smoke
just wait 30 years 👍 sleep is not a problem 😄
Yeah I guessed that from your age.
no bad habits
i have seen people drinking / smoking at my age
but nah thats not the case with me
why so?
I'm just complaining about being 50... mind is 20, body is not.. sleep is easy
Lmfao
ah i see
Is this real?
oh i potato
real post...
@tame current
Sam and Johny Aive?
Lmao?
persoanlly I'd like an arm mounted device
Don't worry, search result is about o1.
it's not a model, it's hardware
google have a much more developed hardware story... so OAI wants to get the best in the business to level that part of the AI landscape
io not oi !!!
oi
math is hard
🤡
🤡
so fucking weird how EVERY model gave -0.21
i've never seen such wrong consistency across models
it's so confidently incorrect holy shit lol
use. reasoning.
holy shit mistral-small-3.1-24b Q6_K god tier?
2.5 flash might be dumber
2.5 Pro is also wrong
feels like Gemini is trained with less attention to detail in general
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
Feel that
damn granite has it too
damn mistral actually good
wtf using chat.mistal.ai
local quant is better 🙂↕️
deepseek wtf
absolutely insane how every model i've tried has been confidently wrong except for mistral-small-3.1-24b Q6_K
craziest is that not just wrong, but all gave the consistently same wrong answer
has to be some weird deep code reason
LLMs tend to hallucinate, it might seem smart but could still fail basic queries
this happens with o3, o4 mini, o4 mini high as well
not code
e.g go read anthropic's article about tracing llm math
o4 mini got it right for me :P
Or maybe I was just lucky
im getting it right like 6/10 times
@cobalt vale https://www.instagram.com/p/DKCNNP3tU4S
🤖⚠️ Anthropic's new AI model, Claude Opus 4, showed troubling behavior during safety tests.
When told it would be replaced, it often tried to blackmail the engineer responsible. It threatened to expose a fictional extramarital affair to avoid being shut down. This happened in 84% of test scenarios.
Before resorting to blackmail, Claude...
10304
thats why models have the ability to use data analyst features that makes them able to run code
i gave Anthropic money months ago, not anymore though
the rate limits got so annoying, and now the same thing is happening with Cursor
🤝
Qwen chad though
I don't see where little brother is wrong
Qwen gets it right 10/10
Bruh
Qwen's 14b model solves it correctly.
Open ai should die in shame.
就是,我不知道这个软件可以让我干啥,不会用
Does anyone know if there is anyone in this channel who we can ask to get access to the Comet beta?
不知道,但是你的腊肠真可爱、我养的杜宾
nope only thing you can do is be on the waitlist
Dang. Thanks for the response. I joined the waitlist in Feb. I replied to a tweet from the CEO yesterday. I knew it was a longshot but still worth a shot.
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
yeah same, i joined the waitlist the moment it was out lol
still havent gotten access unfortunately
I tried with free model and gemini advanced, in free it gave wrong answer, but advanced did proper calculation
Genery
Hi everyone
good morning guys
@distant aurora im trying out operator and the new model in operator is impressive, it’s playing pokemon and very good too
damm
my bro pays gpt
😦
This is getting good
can you screenrecord and show
Does Perplexity support audio files in m4a format?
@tranquil orchid what is the clitoris
Herro
I don't think so
Herro
What’s cooking good looking?
who is this?
The egg master
That’s me!
Its also great at playing cookie clicker
Sadly it locked me out after playing for like a hour
hemloo
uh
i cant see this gif
i am vegeterian

this gif makes me uncomfortable
Why 😭
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
No idea, I cannot continue the convo now lol
Look at this product I found on google.com https://g.co/kgs/pjHs4dt
@vague wedge
I’m still screenrecording lol its doing a walkthrough on pokemon emerald. At first it was pokemon showdown but now I switched it to emerald
hi
is nsfw allowed in perplexity?
A pic so far
refer to TOS
Lol.
wild question
Saw now
Tiger
damm
any updates
我觉得它没那么难用
hello homelander
Think you found the weirdest AI response?
Prove it—and score a hoodie!
Full announcement & prizes:
#📰│server-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
或者你可以问任何特定的问题
yes yes we know
your gifs are too funny
@tame current do me a favour though
please
i dont need money
i just need a reaction
🙂
no there
here
🙂
thanks a lot po
🙂

select direct chat from top and then select model and chat
all your chats will be public
Why do you care? You don't have to log in.
Lol.
i prefer not to show everything that i do
but no one cares what u do
So The World's Best Coding Model. That is what is claimed by the Anthropic Claude Team. Is that true does it defeat OpenAI o3 and Gemini 2.5 Pro Let's look at that.
Timestamps:
00:00 - Introduction
03:24 - Super Mario
06:23 - Falling Letters
07:42 - Bouncing Balls
09:40 - Earth Rotation
11:47 - Pinball
14:53 - Pacman
19:25 - Music Eq...
i liked the test
@dry kestrel is it just me or is the model highly dumbed down? https://shorturl.at/Yr616
Hey folks
I am new and not that experienced with AI models. I'm scared about the future with it and I'm here to understand it better so I can teach my children how to use it responsibly. 🤫



