#general
1 messages · Page 191 of 1
But a.i. could do it in 6 days if he checked 0.1 s per image
Now imagine one day a.i. checks all your data all the time 24/7 365 days
To find some pattern
Whether you'll be dangerous individual or not
One day maybe if you google "how to make b0mb" police will come to your house and arrest you
Same day
what about it is for minecraft if a kid types that
Well a.i. would know it's a kid probably using some pattern recognition
A.i. would know the user who searched that term what his intentions were
Maybe by analysing his previous searches or websites he visited
and what will create that ai and what value will it have?
Control
wait, if the grok 4.1 fast is more smart now, and grok 4.1 is more dumb, the gork 4.1 fast is smarter that grok 4.1 ?
how will shifting opinion be recognized?
Look at China they are using cameras on the streetwalks
And you get a penalty for crossing on red light
They use face recognition and stuff already
us also does that
Probably testing so one day that can go to whole world
It's cheap
Place a camera
Use a data center
Boom
You get free government cash
there wont be cash anymore by then
so what is ur worth will be?
Nothing fun is in the future
A.i. s will replace us
We'll become obsolete to rich people
Why would mcdonalds or wnyone hire an employee that he has to pay monthly or yearly when he can get a robot for cheaper to do same thing without rest
Why would Coca-Cola make a real ad and spend money and time when they can generate a mid-tier ad that people obviously like more than the real one
Like fr people like funny a.i. videos more than serious videos
so how long human can control ai?
but if you think about it
I hope forever
why would there be rich
why?
if there are no customer who will buy those services
if human control ai, u know whats gonna happen. they gonna push to its extreme limits of usability.
Well whole point is control , people with money don't want money anymore... They want control
Power
They're building data centers and everything in order to get rid of us with power
Not to help us
McDonald's would have no customers
they can put themselves in full dive vr to be emperor of the world
A.i. although they're good tool they will be used for bad
I've read in some book before
i want emasterbuild on my side post ASI era
It was made a long time ago
It was talking about two potential paths
When robots come
like what
what do u see my intentions are?
some people says its good.. like trying to make internet more better like more private..
u intention is finding a service that gives temp number for gmail
The transition period (the next 10–20 years) is the danger zone.
If we stick to rigid capitalism while labor value hits zero, we get the Dystopian Scenario.
If we transition to socialism or a post-scarcity framework, we get the UBI/Star Trek Scenario.
The outcome depends less on the technology and more on the politics. If the working class demands safeguards, open-source AI, and wealth redistribution before the ASI era fully arrives, the "Star Trek" future is possible. If the public remains passive while power consolidates, the "Elysium" future is more likely.
gemini
no communism
let me ask anthropic the same
actully Yes.. from the first message that i wrote here.., my plan was someone gonna give me for free afree number just one time code..
This could happen if:
ASI remains privately controlled
No redistribution policies enacted
Transition happens too fast for adaptation
Countervailing possibilities:
ASI creates such abundance that goods become nearly free
Governments implement UBI, wealth taxes, or public AI ownership
Society acts proactively on governance before it's too late
Key point: The outcome depends heavily on who controls ASI and what policies we implement now—it's not inevitable, but the risk is real and the transition period is especially dangerous.
The future could go either way depending on choices made in the next few years.
imagine elon musk controlling ASI
@vast fern imagine if computers replaced humans..
we beacamed g.ays.. and the mach.ines from ga.ys to bo.y and g.irl , so they can reborn..?
did u see the photo i made using gemini-3?
looks plastic
freaky
plastic gnus?
i prefer anime generations, it will be harder, especially swaping characters
gnus?
Basic ahh photos
no i want it to preserve the pose, expression and movements and style
like this, it just did so good wth
if google release frame by frame character swap it would be awesome
how long the heck
by ai?
freaky ass
or...
make her indian
Anybody who works on sora
yes, i just tested the pro one and it much better than the normal previous one, like...it just swap the character shamelessly without change the pose, expression or style
I'm not even allowed to share my images 😭
wdym? nsfw?
bro what
wdym it is not allowed XD
what did i just see
lol
the hell
Y'all ain't seen nothing
i mean i have a program that scrapes all the photos ever posted so
yes
5 i think
in gemini pro
bro the hell
gemini no limits i think
i'm saying nano banana pro
gemini has limit
no way google give us free even free user
only if tpu not available
bro is it free?
But i think it gives better photos
I tried to compare few images fal ai i think gave better results
i tried to send.. but fiscrord didnt allow me.. but the server does.. , thats the censroship u didnt know before!
each genertion is different
is that free
No
how much is it
what the
Similiar pricing to ai studio
how can u afford?
what image u making
Work images
he saying the resolution
School stuff
if you zoom in close, even 4k images still have clearly noise
let me cook something
Yes 4k isn't relevant for nano banana
how many frames in a 2 minutes video
1
Models are trained on lower images than 4k
school or work
Very very important stuff
look at my ohoto 😉 @pine knoll
gta iv?
5
nah it must be 4
look at map
i mean the graphics
this is 5 and that is 4
bro instead of that...i'm curious about your avatar
u want me to post the photo in my avatar?, i tried but diescord didnt make me do it..
Bro y’all need to stop i’m in a public bus
then..., u must better make other people watch with u!!!
really? does discord app has their own filter?
i dont joke..
it just the server bro, you can upload nsfw image to some nsfw channels tho
this gta 4
its disco.rd not the server!!!
dooo uuuunderrstanddm??
bruh your avt is not even a nsfw 18+ image
i know!
thats so blurry
wdym, i have my own server and freely upload nsfw images
ill secure the photo in a link.. and send it to u.. and see if u can send it..
Slow download
I’m having a problem when using the new Banana for image upscaling. It only upscales the face, but it leaves the rest of the figure — the body and the background — blurry, just like in the original image. The only part it improves is the face. I don’t know what to do or how to prevent this
two years ago it was just a mess
It's just your prompting
Try different techniques
You'll eventually find what works
Hahahahahah
it butcher the face if its not in clear view
um...what is the best open source image editting rn?
qwen image edit
rlly?
yes
Hi
Do you really need OSS? Why not just use nano banana flash?
What?
Gemini is light speed compared to GPT 5 thinking, I've had to live with the 5-10 minute thinking processes for months
Is changing the same message now limited? Previously, it could be changed without any problem. It now says "Message cannot be retried". Uh?
Hello, world!
Hello
Hello
@echo aurora will we have new models in the search option? we don't have gpt 5.1 search and gemini 3.0 on it
Hi everyone, could someone tell me if Nano Banana is among the artificial intelligences?
was gemini logo at the bottom always there
or did they just add it recently
Always
If u use app
lighting is oversaturated
At the release it didn't have it, but when they fully released it started to watermarks
This was flagged, but I'll be sure to followup. Does feel like a bit since we've had a new search model added.
what is a search model?
Sry I meant models for search modality - https://lmarena.ai/?chat-modality=search
Yes it is
It's also only in the Gemini app, along with much stricter guardrails. API generates up to 4K without watermarks, and it does images of celebrities etc.
Which is flippin awesome
is there an ai radio that plays good music and you can search and find your vibe? I need it for streaming non-copyright music
LOL
I’ve used nano banana pro for only simple stuff and I’ve used sea dream 4
For simple stuff and I think I’ve got The same results when I tried adding a hat top hat to my Head
Which one is better because I’ve never really tried out nano banana pro
I think Cyanite AI, Harmix AI music search.
Also check out this: https://medium.com/@moneytent/best-ai-music-discovery-tools-transform-how-you-find-your-next-favorite-song-in-2025-caca103c62fa
Get AUDIIO here:
https://audiio.com/jostalking
Welcome to our in-depth review of AUDIIO, the innovative audio library that’s changing the way creators find background music! In this video, we explore how AI technology enhances your search for copyright free music, making it easier than ever to discover the perfect tracks for your projects.
A...
Hello legends! RadioPromo.io is a brand new AI tool that helps musicians find radio stations and contact radio program directors — in real time!
In this video, I’ll show you how it works and how you can use it to pitch your music directly to radios around the world. Whether you're a producer or in a band, this tool gives you access to indus...
Okay, no more announcements today, I promise.
We don't mind them
Good to know
hey @echo aurora you guys never fixed the flux kontext dev license, it is open weights under the Flux Non-Commercial License, not proprietary
better fix it before Flux.2 (krazy-kangaroo) appears next week
What ai is best for coding?
Hmmm I'll followup on this flag. Thanks for the reminder. 
gemini 2 pro and sonnet 4.5 for agentic stuff
Oh boy, that image edit leaderboard sure is a slap in the face for the competition... 😅
Gemini 2.5 he probably means
Gemini 3 is also very good
Anyone here tried antigravity?
Same as VS Code, but with deep Gemini integration. It's pretty cool.
dayum maybe they did nerf it
I mean like, it's literally the same as VS Code.
Does anyone know the model SUNO AI uses?
anyone know a good website to create ai songs? (publishable on spotify or else)
#CLAUDEOPUS4.1 🔛🔝
not even comparable to gemini
Ask him
best logic:
Read index.html
Creating index.html```
@marble wadi what areu doing
deleting index.html
Learning python
you've already lost before you've even started
dont learn anything, just vibe code
🤗
😭
Well i am learning for enjoying it not for professional work i just found its so much fun
im jk
its actually good
dont rely too much on ai
work that brain a bit
Yeah it cant customize to the level we want
Yeah i used the brain and learnt print("hello world") now i will take over 1st place in leetcode
i love vibecoding 😭😂
oh
lol
I too made many apps in google ai studio
Count is 68
Nice but ui design is sloppy and generic watch y component videos to make good ui design through vibe coding
i like this design 🥲
nah that design is good bro dw
dont rely on ai just dont, its meant to be a fun tool
keep doing what ur doing :3
It's worse than nano banana pro
this is the closest any ai model has gotten to somewhat accurately recreating ww2 borders
not bad at all
damn man
def some mistakes but not bad
wow, the history channel watermark at the bottom is wild
so would yall say it hasnt been nerfed,..,
how?
havent used previous checkpoints to tell
you didnt use it at all yesterday?
ah compared to yesterday?
still the same
ahh
so good
well it was dropped yesterday so thats good to hear
maybe res or prompting issues?
i have trash prompts sometimes lolz
i noticed (maybe this is obvious) when it has higher res images to work with it just produces a better output with just better results overall
instruction following wise
that was with lmarena heres w the app
LMAO
Interesting test
Nice
It made this from scratch? If so then that's crazy 🤣
I opened the penn state onefirst and knew that it had to be a picture from last year and I knew exactly whos stance that was on the end
lolz
also cus penn state would never beat ohio state no matter what i wish they would but they are just so incompetent
do it no balls
Man, I didn’t even think of that you’re absolutely right
check the image below
nano banana 2 has google searhc?
Yeah
Dude I found something crazy in sora
Certain prompts in certain images produce 100% copyright training data
Plus voice and beats
And it does so consistently it’s crazy
That means that any movie any TV show out there we could reproduce lol
Or any song or voice
Id be very careful when you upload it doesn’t seem like they have good protection for user or even other data
made by NB Pro, what do you think of this training method?
Bro lol, the grok 4.1 is so bad, it want to eat me 😭😭, I withdraw all my compliments
Carmilla and Laura from Castlevania
why i can't use Banana Pro if i never use it today?
🫠
i only can use Nano Banana 😭
it's paid only on ai studio, use it on lmarena direct chat
Wouldnt let me send it
lmao that was just a good play tho ngl
Mickey
Mouse
It was still
I cant put mickey
Mouse in the same sentence are we serious
well you cant lie any good play ohio state loses to is 'mickey mouse',.,. lmfao
just like how i am with oregon nothin bad
I mean
For the close games sure, games like michigan cheating 3 years isnt mm
Lol
Sometimes nano pro outputs some small gibberish texts. How to fix this? I try re-send the generated image again with the prompt "fix visual glitches and gibberish texts" but it generates the exact same image. Do you guys have any better prompts?
depends on the situation, if that text is really important for you then do this
Circle that area in red, and clearly state what you want it to write there
Inside "" and prompt nb2 for it.
Moreover if you just want coherent and readable text then just circle that area like first step and prompt it like this
Priortise: making the text in marked area coherent and clearly readable instead of random gibberish
short animation made with nano banana pro
Ciptakan vidio dari gambar ini background dan suara masih sama. Pria tersebut menggunakan pakaian yang sama pria tersebut berkata.“Assalamualaikum warahmatullahi wabarakatuh.
Saya ingin menyampaikan kabar gembira untuk Anda hari ini. Setelah dilakukan pengecekan dan verifikasi data, Alhamdulillah kami dengan resmi mengumumkan bahwa Anda terpilih sebagai penerima hadiah sebesar lima puluh juta rupiah
Thats actually insane
Her shoes 😭, still that's really great but wdym made by Nano Banana 2 were you generating alot of images or what?
Is that real or ai?
Before it made the image Nano Banana Pro have a thinking method before making images so it solves the question you gave it and then make image about it on the board, it's pretty insane. Example:
You can only see it on API I think
hello
Nano banana pro is insane at creating anime images
Hi
Have you gaved it images or it made itself without images
Made itself
Anime art style, split composition. On the left and right, Naruto and Sasuke looking god-tier powerful with glowing chakra avatars and intense lighting. In the center, Sakura Haruno looking scared and hesitant, standing still with no aura. High contrast, dramatic shadows, 4k resolution, Naruto Shippuden art style, --ar 16:9
Dang that's sick
This is the prompt
Dang
Ai is getting out of hand
That's the best poster wth
Graphic designers are done
Make it do a poster for itachi or smth
You read me or something
I mean yeah they are kinda done but still it's not that good so
No bro figma is already included this to their ai
It's so good at ai
And it's so good at youtube thumbnail
Literally half YouTubers are using ai thumbnails
Dang
hello
@everyone need a editor
Nano Banana Pro is finally here — and it’s the biggest upgrade Higgsfield has ever launched. Powered by Google’s new Gemini 3 Pro Image engine, this model is built for control, clarity, and structure… and yes, it now generates in true 4K.
In this video, I test Nano Banana Pro live on Higgsfield.ai and break down everything that makes it...
Dam why is everyone making ads about Higgsfield??
yeah i alrd know
try them out!
seems to me like nano banana 1
not the pro one
and it has queue
https://flowith.io/ is way better
Where Ideas Flow:Interact with the world's most powerful AI in a way from the future flowith is your AI Creation Workspace that transforms knowledge. Through innovative interaction, it allows you to collaborate smoothly with AI, with ideas flowing like a vibrant spring.
hello
and bye
higgsfield is fast and their tools are endless
it processes for 30sec max
can i use search + gen image?
can i upload doc?
can i get 4k for free?
Why when i create nano banana pro 4k sometimes 15.5mb to 19.5mb sometimes 8.6mb(Please answer me )
They paying them for sure
That's how image works
Same with videos if you download a movie in 4k it will be 6GB or smth
how wexactly
It's overloaded that's why
Just wait about 10 minutes and try again
I hope lmaeren update the resolution for nano banana pro 4k or 2k is fine
That will cost alot it's very expensive
2k resolution same price as 1k
yes 1o generation everyday
I think they won't add 2k or 4k but you can try it for free on aistudio not now wait until they release it for free because they are trying to improve their cloud or smth
Alot of people gonna over load it
Because there are many users on it and now overloaded
aistudio Free, how?
antigravity is really making google look bad
I said right now it's not free
the IDE is so bad
I don't care as long free models
its not bad for being free
Yep
I have 20 pro account, use it in the same time
Spread this around if you are tired of having constant bugs and feedback ignored https://discord.com/channels/1340554757349179412/1441771173020045473
is ts profile serious gng 🥀
Yes 😔
dawg
wait charlie kirk is dead?????
-# ⓘ I hope lmaeren update Nano banana pro with search feature more](https://discord.com/terms/local-laws)
-# ⓘ This message is not available in your region • Learn more
-# ⓘ This message is not available in your region • Learn more
<@&1349916362595635286>
Dont click its a scam that will steal ur account and bank account
No wayyy really?? Look soo legit 🤣
Some people can fall
@neon idol Thanks for your report!
Ur welcome
First time seeing some mod other than pineapple
Why does it say this message isn’t available in my region?
Me too
?? Why this message is don't available in my region? 😭
-# _ original message was deleted _
-# ⓘ When will LMArena add PDF Uploading?(https://discord.com/terms/local-laws)
last thursday possibly
I need this 😭 have 6 month I'm feedback this
Bruh, I said something sus again?
Hey is there a limit to video generation on grok.com ? I did a bunch yesterday and now it gets stuck at 0% but I get no error message.
Melon musk said unlimited, but idk
Melon Musk 😭😭
For free or for SuperGrok? Either way I believe there is a limit (when you access from “imagine). Grok claimed to have a dynamic limit depending on traffic or else. And cooldown sets every 24 hours since your first generation.
That’s my take from what I searched last month.
It is quite unanimous Gemini wins
Waiting opus 4.5 🤤
It made the most precise calculations of my giant towel flipping question
Gave me an answer of physically non-absurd towel to fliip
It says I'm on the free plan (I have a basic x.com account).
This is what I'm seeing in 2 different browsers.
Correction: wringing
If it stuck at 0%, most likely you hit its limit for the past 24 hours. In mobile app, it sometimes tells you limit reached/come back tomorrow.
I do not know physics, much less applied physics enough but I know who is non-credible
I use dumb questions like: For a cable of plausible diameters, what is the longest cable permissible physically to whip in the atmosphere, even I know it is absolutely going to reach hypersonic speed?
Uhh
Towel? 😺
Hi
wdym
looks like old model
Mokou talked about towels - so I suggested which kind of towel to use. 😸
Gemini 3 pro nano banana
dexter MORGAN
Actually some do, those who are found in somewhat colder climates got more fur.
Someone had get mercury? It was updated in 06/11
its surprinsigly good at making car interiors right
I'm more impressed with AI's that are able to create things that do not exist.
In this case most AI's made a mess of the undercarriage, as they did 'not know' what to make out of the looks of the seed image.
Clanker cracker
this is not a prank right?
Anyone want sora 2 invite code
we're still not there yet
Gemini 3 Pro is insanely good at coding bro
lies
yes it is
there is no gpt image 2
i think its a prank
It Is a prank
i meant lua ui designing
no
startergui
there's a startergui on rblx studio
its insane for roblocks actually
however gpt 5.1 high generated a better terrain from command bar than gemini 3 did
yeah that's true
gemini 3 is mostly good for ui designing
bruh, first BATTLE input of the day and ReCAPTCHA🤣
Audrey Kathleen Hepburn (née Ruston; 4 May 1929 – 20 January 1993) was a British actress.
Recognised as a film and fashion icon, she was ranked by the American Film Institute as the third-greatest female screen legend from the Classical Hollywood cinema. She was inducted into the International Best Dressed Hall of Fame and is one of a few en...
gif's from the movie Breakfast at Tiffany, i have a few jewelry from that brand
actually... no 😅
https://www.mellerio.fr/pages/au-fil-du-temps
i consider this brand the best since they managed to survive the French Revolution
show example
I love the battlemode for comparing output. 100% fair to all.
my device cannot do it..
I have a question: which is better, Grok 4.1 or Gemini 3.0?
May both AI technologies take a big leap forward
3.0 R$%# the grok 4.1
grok 4.1 better
lol bro this AI want harras me
I don't know about the fast, the normal grok 4.1 said it loved me and I said I loved it lol
what is your use case
I currently use my own cloude API for automation.
But I don't know whether to switch to which of them
I used deepsek before.
It was good, but I think Claoude was a little better for that.
What AI do you use?
I'd be happy to help you with that.
What exactly do you do in it?
Nice
Do you write your prompts yourself or does some AI create them for you?
A good start
Myself
what is the best ai tool to learn
gemini app, guied learning
I don't tested with gemini 3.0 but even it 2.5 pro is the better AI to learn, my top 2 is deepseek v3.2 (idk why)
Lemme test chatgpt 5.1 vs Gemini 3 Pro for a simple prompt
but its so limited
hmm, Idk another better
after 6 prompts they will hit me with rate limit
ahhhh, I have a pro account I forget this 🫣
hmmm free? yupp not, lmarena not, but have the flash 2.5 bro
dynamic mode is much better
have you tried learning anything ?
@Gork
I use to study pharmacy and prompt, (I know I need to use it to learn english lol)
grok is little retarted it gives wrong facts
I'm testing to make post now, but in slides (powerpoint) and docs (word) yep, the gemini 3.0 is better
Is copilot good?
Ooh
Aight tell me when ur doing testing
Done*
but I don't fell OMMGGGG IS AGI NOW
wheree?
What AI should I use for prompt?
reaaly??
is giving for me sooo headache o-o
so its with ur hands?
or u are headpain beacuse of agi isnt real?
like i mean by "not real" is "didnt released yet"
no, because it don't know to make the baseboard of my html posts
the rest is perfect, better of me making in canvas, but it don't know how to make the baseboard 🙃
which one? do u mean the semi-agi "Grok-4.1"?
Idk if I use it agentic use I don't have this issue but nah, is away from agi today
brah , let me do it for u.. give me what do u want.. ill do it for anything in exchange..
how so? 🙃
giveme the ideas or the prompt, and ill be ur new agentic..
u could do it in the private btw..
one free trail , and the other is for exchange , like instead of u pa.ying for ai..
hell no, the gemini 3.0 is giving headache, but the opus 4.1 and gpt 5.1 is killing me, the gemini 3.0 is obviously better
So, I accidently hit retry in LMarena when I meant to save the image. Is there a way to see old generations?
I'm busy now, and my work I make in my lenguage too, but thx
Ok
@fiery gull get me ur thing u want to make.. like pretend that im ai..
don't have the history chat in the side?
np.. ill also reply in ur language..
the memory issues bruh
while the on eon ai studio can
why u dont trus.t me? is there is something wrong or i forgot?
noo, I just busy now
I jump to discord when the Ai is generating
so whats u are working on?
in posts now
I'm using gemini 3.0 to it
I fell I make this more fast in canvas lol, but I don't care
mamilex what language od you speak
I have the chat history on the left side, but I don't know how to see the old pics that I generated. I just assumed they would be gone forever when I hit retry.
dude.. why not let me do it? im great at prompting.. , just give me the error , or whats wrong.. or just give me the html code.. or part of it.. i can fix it right now..
i have jailbre.ked gemini 3.0 in the first day-released..
ahhhh, Idk about this, call the @echo aurora
bruhh okay
OMG, Great
Do y'all think this is good for Gemini 3's coloring (nano banana pro)
omg, it perfect ;-;
sukunas jawline can cut my finger
wait wdym this is colored by gemini?
thats so good
yeah the prompt is "Make this highly detailed colored. As if it was in the anime but it was drawn by MAPPA. (must be STUDIO level quality.)" :)
Mhm
No, I don't think so. Tripo AI is good at AI modelling from image refs from what I've heard.
would be crazy
Modellers would love that
i don't understand this typa nerd talk 😔
is there a way to create a video from an image? can anyone help?
;-;
@floral crown Please head to #1397655624103493813 for a detailed guide on how to use the bot
yo bro. You should not remind me of that image
It gives me legit ptsd
500 cigarettes
I don't have tiktok tbh
so I don't know the reference
This shirt isnt tight enough 😡
I just wanted to see miku and teto in star trek
luck
Nanobanana cant make a football shirt this tight
guys, please rate my ui 🙏
https://llmplayground.net
Create and talk to AI characters
looks like any other ai website ever made
honestly? thanks 🙂
thanks, noted for next update ❤️
helps a lot
guys, please rate my ui 🙏
https://Google.it
Search the world's information, including webpages, images, videos and more. Google has many special features to help you find exactly what you're looking for.
also this is too dark, bad for readability
red reset button is ugly
Is that link sus?
ah, thanks 😊
❤️
both noted
and seen
so will update those, will take one variable each
have a nice day for now, and thanks a lot again for the design tips 😊 ❤️
If I make a LMarena account will i be less prompted for captcha?
This is so unusable I am constantly being prompted of captcha
No, even if you create an account, the new captcha doesn't work. It continues to ask for confirmation and fails the checks. LMarena is completely unusable, but the development team doesn't care. They've opened threads in the peer feedback section to ignore user responses.
works fine for me tho
does lmarena have api?
no
No
sad
To get API get it from Google, OpenAi, Claude, Grok and etc...
ik
thanks btw
sg oe
<@&1349916362595635286> user id: 1158181101798436984
Promoting scams
Nvm it's gone
Thank you. Banned
the AiStudio has api free
Hello. Please note that the warning was not sent because of that. You can give as much feedback as you want as long as you are respectful as indicated in the server rules. Thank you
don't work
Hey guys, seedream is 4k, is there any chance nano banana pro will have 4k in lmarena?
Yes but it kinda have limits
Not confirmed and probably not because 4k cost more so they will stick as 1k, wait until Gemini Nano Banana Pro be free on aistudio then you gonna change the resolution for free. Right now it's not Free you need A Paid API. (not confirmed tho)
Wb now?
Tried the 4k one its crazy good
now is woking, I'll see it
It's fun
What's that
his eyes 😭
?
see this post from gemini 3.0 to instagram is INSANE
I'm using it now lol, goodbye canvas
Wait i don't understand? Is this post ai generated or something hahah
gemini 3.0 pro with html
What like working websites? Or just files like we have in build?
If it was working websites with my domain names it'd be dope
idk, I just making post it gemini 3.0 pro
I don't test the nano banana pro yet
It's so scary good
Like
First time i did good prompt
I cried
How good it looked
Like i saw this and my eyes were full
This was my first good image with the pro
nano banana pro?
How r u so fast
Can we like change the resolution or something similar to that to make the image more clearer?
Yes in aistudio or other websites using their API
On this lmarena site?
You can set 4k which is expensive i suggest 2k since i doubt their model got trained on 4k images
yes
Broo stahpp
Not on lmarena's
On fal.ai for example
But it costs
0.15 per image (1k,2k), for 4k it's 0.30 i think
What
Kissin
Lmao
Hii
I couldn't get the transition for the laptop to emergency vehicle because of "content violation"
how come no content violation for the figures
lol
hello!!
Hi
What is your favorite kasane teto song?
hard to choice one but I'll go with ochame kinou
I'll see, but instrumental lol
wdym 😵💫
I don't like voices
I have a big playlist of vocaloid song but instrumental
is it public
I'd like to listen that
No, only in my phone downloaded
Kikuo
that's a very good choice imo
i can still hum some of their songs probably xd
The 3d song from kikuo is insane 🤤
i use my laptop for using lmarena for ai picture and it show this "Something went wrong while generating the response. Please try again."
F5 and recaptha again
i did but i didnt work
Hmmm, try to use another model 🙃
The opus 4.1 still good
for what?
Overall, but the gemini is better ;-;
yeah thats what i was about to say
it was better at coding before
but now its not
is livebench paid by openai?
Gpt still top 1?
top 2.... by like nothing
it says that gemini 2.5 pro is better at coding than gemini 3
i dont think its a good benchmark
Sonnet 4 above 4.5? Okay... But gpt 5 chat? Hell no
4 is definetely not above 4.5 lol
what benchmark is that
livebench
Live bench coding
im honestly extremely surprised how close grok 4.1 is to gemini 3
why Gemini 3 is the best almost on everything according to lmarena leaderboards
on lmarnea
is it THAT good
because it is
yes
I wait this, because the grok 4.1 is pre trained to lmarena
wdym pre trained to lmarena?
pretty insane
Like the llama 4
grok is so unfiltered i love it
It harras me
whats even more insane is that its free from google
😭
well lmarena funding makes all models free except gpt 5 pro
and image models still not really free
Likely the gemini 3.0 os better that gpt 5 pro 🙃
tried to make a flappy bird game with both sonnet and gemini
Too much captchas with the launch of nanobanana pro in lmarena
geminis art were a lot closer to original game and the gameplay was a lot better
sonnet did an alright job but like gemini is on another level
also are the ais not able to make more than one file at once while coding or something
Has so many bots now
Yep, At first I thought the Sonnet 4.5 was better, but that was because I didn't know how to use Gemini 3.0 🥀
thats what free 14 cents a prompt gets you
I understand but is there another less annoying way of detecting bots?
Ask to devs
Maybe a mandatory login?
I failed the choose all the bycicles one 😔
wouldn't fix much imo
Very difficult
captcha detects you a lot more when you type things fast
which is quite annoying i agree
They did this with mandatory Google login, but people started hating on it, talking about "privacy" 🥀
lwk valid complaint tho
mandatory login is not a permanent solution and
it has its own problems like privacy
In my opinion, I would prefer a mandatory login rather than having to complete a verification every 30 seconds. 🥀
yea but what i am trying to say is, if the bots are so much of a problem that captcha shows up every 30 seconds
mandatory login is not helping our case a lot (i think)
showing captcha every 2 minute instead of 30 seconds would still be frustrating 😭
I’m fine with the captcha being just the one where you press the button to confirm that I’m not a robot, and that’s it. I don’t want the other ones where you have to select every square that has a train or a bicycle or something. Because sometimes just one pixel of, for example, the bicycle is in a square and the captcha fails because you didn’t select the square that had that miserable pixel of the bicycle in it. And it counts as if you didn’t select all the squares that had the bicycle
mind if i ask in what mode and what tasks?like coding, writing, or image?
Mostly image generation with any model not only gemini 3 pro
How ican fix this? gemini is not showing old image generations
can u guys guess what image generator made this
Trick question, it’s a photograph
do yall think the leap from gemini 2.5 pro to gemini 3 pro is very big or small?
gemini 2.5 pro was already so good
so medium
oh
ahead of its time
lol
whoever voted downgrade must be trolling
hi guys, how do i increase my daily limit for video generation in the video arena. please help 😄
2.5 pro is small, the servers from google need to handle 500mi of users, the 2.5 pro must be small, and 3 pro don't is more expensive for google because they removed the 2.5 pro in same time, and the 3.0 pro need to be small to handle the servers, I think the 3.0 pro is 3x bigger that 2.5 pro, buttttt it is 3x faster that 2.5 pro, you get it?
Ahh you talking about smart, I tought is the size of model lol
Hello Everyone
salutations
What are the best third party providers that have generous free credits for Nano banana pro?
idk what lovart are doing but the output is way better than any platform
What about lmarena?
Hi
Create more accounts or just use grok imagine, it is free and have audio
the quality or limit?
Both
quality same as in gemini app (1k)
here is an example
loveart
gemini app
loveart always outputs more details+graphs etc...
i need to extract loveart system prompt
i can try to fetch it for ya
if you want to
yea pls
dont tell me kimi k3 will be released before deepseek v4
Kimi vision?
/imagine prompt: Medium: hyper-realistic 3D medical illustration.
Subject: a man eating a pomegranate, shown in detailed anatomical cutaway revealing brain, lungs, heart, stomach, and intestines; realistic skin texture, transparent chest view, exposed organs with medical accuracy.
Emotion: neutral, focused.
Lighting: soft studio lighting.
Scene: front-facing torso view against plain light background; clinical, educational atmosphere.
Style: medical visualization, detailed textures, anatomical accuracy, clean color palette --ar 3:4
Not even close lol
I am a large language model, trained by Anthropic.
Therr is no command dude
Yes it got nerfed alot, I feel like Gemini 2.5 Pro is great at coding
I keep seeing errors on my project everytime dude Gemini 3 is just weird
User id: 144044152467541#### (banned)
Scam or smth
<@&1349916362595635286>

lmarena ranking system explained
not bad ig
i need to check to see if those formulas are right
sorry i couldn't do it
no app in mobile provides a proper browser console
i miss kiwi browser
wonder if it still works
heyyy
i extracted it but doesnt seem to output the same as in their website
:/
that's pretty weird
what kinda sorcery do they even do atp
btw what's the prompt
Hi
how u doing mnan
hi
how can the text be generated so well? mine is just not gibberish on my NB pro
hello
What happened
U need to mention infographic in prompt
Sommeon was promoting telegram scam thing but pineapple already banned him
I'm great 👍
Then he deserved it
Still can't use nano banana pro
where r u from???
Hello, , glad to be here
was anyone using Flowith ? up until a few hours i could generate things on nano banana but it doesn't seem to work anymore
Your welcome
