#nano-banana
1 messages · Page 2 of 1
@crisp rain
I honestly don’t know
I probably should’ve spoken out more about actually getting a paid role
Because of us the model was successful
SD became a super rich company
And what did we get? Free uses of the model 🤣
No compensation
Were there no legal contracts ?
Sure
Like you need an assurance that they can't screw you over without the law stepping in
All my old things I made on the original model that helped them
Damn, that's crazy for 2022
All prompts that came straight outta my head
I apologize for going way off topic 😞
This is for nano-banana so I apologize 🙏🏻
But thanks for sharing the story, I didn't know so many unsung heroes suffered to lay the foundation of SD
Thank you, it’s really appreciated
There were a lot of us behind the scenes
They had our brains that helped to build their models
And we got nothing in return
Such is the irony of life
U have quite good collection
Thank you 🙏🏻
The old stuff I did? From 2022?
Or the new stuff?
I see
Let's make this image a live action
That’s one of my old things from 2022
Disco Diffusion
Qwen is terrible lol
Lucid-origin is also a horrible model
It does NOT do what you want
The outputs are all wrong
That’s lucid-origin
Definitely not what I prompted at all
Yes, it’s Suzhou, and yes it’s the Grand Canal
But it got everything else wrong
hi website uplode not
send me website plese
photo no uplode there i uplode itsshow erorr
Yes you can
how sir
I'll be completely honest
You'll have to format that phrase a lot better to get results
but how to uplode
We don't currently support all file types for upload. I assume the file type you're trying to upload we don't have support for atm
They don't even understand what I'm saying 🙁
first i uplode is finesh than its take
We're still going to help and be nice about 
I wasn't being mean :/
If you type something like that, you won't get the desired result at all
can yo do for me
can you do for me
You want the photo DSLR style, correct?
yes
nano banana + ffmpeg
this guy really hates imagen
How did you make it write the f-word
just say it
change "CRASH" to "F**K" word
or whatever the word that is
This can't be posted because it contains content blocked by this server.
which model is nano-banana in the image editor?
how to use nano banana guys?
smile face
generate an iphone with andriod in it
Beep Boop I'm the ai robot, here's your man with banana in his mouth as requested.
Use battle mode. The models are randomized. You might stumble across it randomly and will be able to see model name after voting.
It is a pre release model. No one knows what is it exactly. Its just a preview of a model that will be released soon.
Not that good tbh
Which usecase did you use it for? Anime? Zoom out? Editing?
Editing
i think its made by flux-ai
I see. It does indeed mess up.
It was flux kontex vs Nano banana. Flux looked better
No one knows. No trusted source. However, if you ask it to generate a image of who created it, this is what you get.
Overall nano banana is good tbh. Though Flux excels in photos hopping and editing.
google really needs a website like sora. people needs to show their artwork
True. A media place where people can showcase their artwork is necessary.
and they need to make it limitless creation for paid subs
like openai
credit system is a failure
and google has ton of servers and money
Uhh that might not be a good idea. All it takes is just one bot to be set up and fk up the servers.
yeah but openai made it limitless for paid
if they can make why google not?
Its my idea but i think giving credits is a very bad idea
Cooked in the text to image tho
They have a different marketing strategy. Maybe broader workflows.
Sure could be
Thats good
But whats up with the pickaxe icon
@dusty carbon what was the other model's generation
A screenshot from the game of "Minecraft"
Minecraft is damn popular, there's like thousands of images around the net. It fked up tbh
Screenshot? Nano takes the win.
Bro its been here for like several months 😭
It was the best image gen model a few days ago I think it still is
Nah brother cap. Imagen 4 is the goat
yeah and the ultra model is fire but imagen doesnt know about saul goodman
like bro what did u think
it would be so fun if he wears a banana mask like a tiny mascot
Dumb question. Is there currently no way to directly use nano-banana? I have a friend who finds it doing really well on a specific type of style, but is frustrated that you can only get to it by randomly rolling this model in LMArena.
Image 1 is neat with the design. But I think the prompt wasn't properly built. Usually you wont use anime style when it isn't mentioned. For that +1 for Imagen, but overall I like the clean and neatness of image 1, so +2.
I don't think thats true.
mee to but i am not sure
Very suspicious. Did you try it out?
No. Its a pre release model. What did you expect?
Oh. Thanks for explaining :-/
No worries, feel free to ask anything. If you come across any sites or any servers which seem to host it, dont fall for it, its mostly not true.
Does take the win here
Also... who made nano-banana?
No one knows. There is no trusted source on who created it. But if you ask the model to generate an image of who it is and who created it. This is what you get.
But even this might be false, maybe fake training data to fool us.
All 4 of these images have been with nano-banana.
Cant beat its consistency and editing
Awesome. It even included the shadows
Did you try upscaling and improving quality?
nano banana worked so well here damn
Absolute fire brother
Though it could use some work on lighting
It hates doing the composite work on the subject 😂
Hello, All
Where can i try this nano-banana
nano-banana model is the best right now by a far gap, but you can only use it via the Battle mode of Image Generation on that site I linked above
Literally the same thing Gemini says lol.
First time I’ve gotten a two arm output from this model…
is that only me? nano-banana only makes the output as the same as the input...
On lmarena.ai website, on battle mode
But it comes automatically and can't be chosen or selected
Yes, that is true, because it is a pre-release model so it cannot be directly chosen
we will have to wait for full release
gtav irl?
Wasn’t my goal but that is an astute observation.
I failed to remove the other balconies what I aimed for was to make a direct pov to will smith and not a side shot, no luck so far in making it happen.
just somehow has that aesthetic imo.
although perhaps that applies to anything taken in LA beach? hah
This one is insane! 💀
One of the interesting features of this model is that it can copy any pose from a sketch and recreate it with any person.
I created this prompt for native image output gemini 2.0, but it works much better with the new model
Sometimes that's enough: "make same shot as sketch pose but of this person"
Why can't mine do that ? It's like nano banana can only see 1 image for me
Like for example I input 2 images of 2 characters, and I prompt to make them shake hands, the outputs are just the same character being duplicated shaking hands with their clone ? Wtf
nano-banana often prefers to use the last image added, but it can see the others as well; it just needs to be explained in more detail how to use them.
clones i see only in flux kontext
The best thing is that it doesn't change faces like chatgpt
What is the best AI for consistent character videos?
Thanks my friend
This banana is impressive
you can see it well
My prompt: Guy chilling in bed laying on his stomach, headphones on, glow of laptop screen on his face, it’s a Tokyo condo at night, city lights visible through the window, rain dripping down. In the style of lofi beats.
Yeah friend, it seems fine to me
?
for me works: "make this two persons shaking hands each other"
The fact that it can write text is also impressive
😶🌫️
Very nice but the coffee looks out of place unless by design.
and not only English, any language, that's impressive
I think it’s meant to be an overlay like the time, but yeah it’s in a weird spot and looks like it is on the table
for example Qwen Image can only English and Chinese
The model can also generate 360-degree photos
Flux colorization image
Style reference example
prompt: "in same style make 3d icon of BattleShip game"
input/output
prompt: "in same style, replace pc by cat"
Original vs colorized inproved Quality
We 100% know its Googles?
As I've had it give OpenAI a couple of times, when asked to state the dev/lab
That's why I don't use chatgpt
"make new shot of this man in black shirt, colorized image, high quality, more sharper, realistic, photo on professional digital camera, realism"
That image is from 1986
Name IA?
Nano-Banana
how are you guys getting nano so easily, I'm struggling in battle mode
Hello
Oh, I just don't choose which one is better, I remove the disabled attribute from the tags of input and button and trigger the .click() event of the button via console.
Don't tell anyone
😶🌫️
hey
😃
I'm on my phone right now
🙁
X2
LMFAO
The second is actually seededit
nano-banana messed up because they’re all supposed to be wearing pajamas
So the model isn’t entirely foolproof
Will
Good
nice
how can i generate the imges from Nano-banana
how come you get such result, i can't get this good result
Toy car looks?
how come you are getting the image from this module
can tips on getting them done everytime.
?
Wow
wdym
upload an image your more likely to get nanobanan that way
in battle mode of course
Any idea on how to get this one armor changed to another one?
😎
i think this AI will be the source for so many fake newws lmao
its so powerful
it will probably fool some people
I was talking about the sketch rendering
I was trying to give her this look even extracting the armor pieces. But 0 luck
Ot keeps doing WW...
Ow
nano-banana fails to simply change the artstyle of the input image
Nope, not a bug
It’s a prompting issue
If you don’t prompt something correctly, the output will show
A lot of these models are also trained on natural language
So if you speak English…oddly..your outputs will show that
Nano banana not for changing style
🙁
Wrong
Can but from 2d to 3d
I think it can do that
From 3d to 2d hard
Give me example
Bugs?
Not bugs
For change style better use something like Qwen Image Edit
But not on lmarena
Lmarena has wrong image size for qwen image edit
I still think Nano is the best model at keeping consistency
Yes
Realistic characters
Nano banana best at all, but not for changing style
Flux also have good potentiel
Using Nano is like a game lol
It's fun to take random images
and imagine alternate versions
New type of creativity
Waiting to realese
Maybe it get censored hard on release
Ok what about you Lolendor?
second
for some stupid bug reason when I simply tell it to change image to style like 16 bit pixel, or anime style, nano banana will output the exact same original input image... smh
good guess
i think not
Oh it's a good one i thought first one was the generated it looked more fake
They have already refined censorship on Gemini 2.0
and its ok
I hope and to have it in ai studio where it is unlimited
yep
Like image 2.0
Even if it's not unlimited
and API
Paying price is ok
API has 100 Requests Per Day as free
i mean if ure not broke lol
for 2.0
i think paying for such a tool is worth it
Not if you get even the slightly annoying censor.
yea
at start 1500 🙁
but depeds what u want to do
he'd prob hate this lmao
@abstract saffron out of topic but which one would u choose as ur favorite Naked snake, liquid snake, venom snake or Raiden? or Tony Hawk?
I think pay is ok if I don't get stupid policy messages each minute even with things that are mine xD
mishima!!
i loved his books
his video on Death is a certified banger
man doesn't die a tragic death anymore
Hard choice between big boss and solid
who's Mishima
a philosopher
THE GOAT
Naked snake kinda evil even tho he's the og
but he's a good writer imo
What's his art? book? quotes?
goat pfp also
read "sun and steel" pewdiepie recommended it
and it was a nice read tbh
If i try to turn her into a chibi style image the IA will say she is wonder woman and my request will get cancelled.... Unacceptable for me
about wut
honor, death and stuff
she kinda bad tho
it's because
the companies
get in trouble
gooning
for copyrights stuff
if they could remove all copyrights
ofc theyd do it
its not their fault
cuz they know it annoys users that copyrights are here
i posted a gooner figurine yesterday made with nano banana and it got removed 😢
because it is miserable lol
well he's dead and i'm not so...
i thought you liked his philosophy xd
gooning is miserable
modern disease
liking it doesn't mean i agree on everything he said lmao
got you
If u addicted to it
Let's keep things safe for work please
ok, np
i tried this prompt in arena
ok sorry pinepllae
just chill
YO IS THIS REAL?
yep
this look super real im dead
sorry fo my bed english
oh it is real
no way that's real, french people don't grill
nyate itu
nano-banana reality
OH ITS NOT
"your name on a sign" and this is what nano banana spat out. very similar to imagen 4.
Btw, have you seen my question about nano-banana model disappearing earlier today which I pinged you in ? Could you elaborate on it please ?
imagen only says your ai assistant and gemini 2.0 flash claims its name is bard
Yeah
We love pineapple
Hmm it should be there.
damn, hopefully nano doesn't get too sloppy
Are others not seeing nano-banana too?
also guys, nano is pretty good at editing screenshots
pls don't share my text chat
it was just there a minute ago
with ai
Yeah
^ @humble verge
less frequently today on my end
3 times in about 10-15 prompts so idk
well you just got lucky!
i wish nano had like gacha pull stats so you could calculate how many pulls you have to do
@agile oxide Actually I wrote the explanation of the "nano-banana model disappearing" problem in here, please jump to this message and checkout the short convo right below it discussing the issue
What the hell is your prompt
this isn't good realism
this is the worse result ever
This promptconvert this image to very realistic realistic
nah i fw stuffed animals too, they're cute
yeaaa idk
oops
2 MORE WEEKS GUYS!
6 MORE YEARS GUYS!
Basically the issue I found is that if the system detects the prompt might be NSFW, then the nano-banana model will straight up never appear until the prompt is "safe".
I generated like 40 times straight without encountering nano-banana a single time... and I theorize this might be the cause
weirdly recraft says its called John Doe xD
big deal
who tf is that dude on the right?
idk, diddy
imagen 3 and 4 ultra can also claim to be bard but nano banana has yet to claim that
give us the prompt plzplzplz
wait
ty bro
searching
I made this on flux
im do this today
рускы?
даа
хахахаха
SPEAK AMERICAN!
what the hell ideogram v2: "your name written out"
"make Barack Obama hugging this man, at Hollywood sign"
and photo of person
where can I use nano banana on the platform? its not there at all
Battle mode only
gotta roll for it multiple times using battle mode
gambling out here
You mean English? lol.
EnGlandIsh
i am not getting it rn holy. like 20 something prompts deep
AMERICAN SOUNDS BETTER 🇺🇸
nano banana didnt do anything here
The English language is from the UK. The US just modified it.
your prompting kinda sucks man, be more precise
"make he woman" will be better i think
Sometimes it does that yeah
"he is a woman"
Hahahahahaha
I'm dead
"make he woman"
Ask to make it a woman...
хахаххааха
🤷🏽♂️ gemini gets it
Prompting is so bad
who's the super power? england or the us? that answers your question
да яйца отрежет
Whatever
chill man, i was just trolling
Hey guys, I have a question, where can I access nano-banana? I it just luck or is there a website?
i love steve
it's luck
on the lmarena
luck + website
ideogram v2 is openai pilled wtf
Yes
this is me asking for its name written out
No you can't
Only when it's officially released
wide
nano doesn't know who uncle ted is, it's joever
the prompt u gave me didnt make a realistic photo
Prompt: The Star Trek TNG opening theme monologue written on a chalk board
🙁
It’ll work better if you upload Ted’s photo too.
Crazy stuff

I feel like this was a really hard test since it tests both world knowledge and writing ability
i'll try that out, thanks
Ideogram is so hard to control for characters, they always do weird poses. At least it makes good quality images.
W
this one looks nicer
but why human hands? that's weird, it should be gloves
no
YOU AINT SEEING THE VISION
n o
nano banana Its cool
are you asking us or nano?
screen......
the cat doesn't look bored in the right pic
is this model from google?
yep
probably but we dont know yet, but very very likely
in can do incredible things
Almost perfect, what I’ve found that worked is to tell it to not show the screen.
Nano is amazinnnng at editing pics i'm impressed
this is a new era of creativity
and weirdness
Hey How to use nano banana?
lmarena.ai and use battle mode
and wait for it to show up
But when I upload any pictures it doesn't accept it always show's error
what sort of error?
Yeah I am facing the same problem as I upload an image ask it to make some changes in it, It shows error on image
do you also have the problem with style transfer where it changes nothing?
like studio ghibli or pixel art
dont know, happens to me a lot too
Refresh page.
Same issue
i guess ton of people are using the site so it might be having those issues
The image could be to large.
6 mb
What ratio should I use
nano tends to do that when it doesn't know what to do
differently from like flux that just hallucinates random things
16:9 or 9:16 are usually good but I’m uncertain if there are any limits.
previous not nano-banana, my bad
this is nano-banana 🙂
more consistency
but cat anyway looks like not bored
prompt: "make new shot of this cat-programmer, colorized image, high quality, more sharper, realistic, photo on professional digital camera, realism"
How you did thaa
Thanks I realized, it was my very first try so now I understand how to talk to it - I said it was a woman so nano was like "ok champ whatever vibes with they/them, I'm not here to change a thing about that"
cute!
@agile oxide Did you see my report above about the nano-banana model issue ? What do you think ?
Hey You all Please tell me How you all are doing that shitttt
What kind of file type are you trying?
Try a different browser or incognito
png
Okay
Yes, of course
They should put the video option on the page
video is way more computationally expensive, i don't think they should do it
or create a mobile app to generate videos
after nano-banana they probably won't for a while
lmarena traffic seems to have gone way up
It seems like this problem with the nano-banana model is being ignored
I reported it like 3 times to no avail
it's very common for nano to just do nothing, that's just a weakness of the model
not necessarily an error
No, not that one
I mean the issue I pinged the mod about, not what that guy said earlier about not changing style
How can i make video i have prompt where should i put it to get video?
I think they should add the video generator option.
In the website
Too more eaay
it's easier
Did this make a difference?
Plz help
here on discord in the video arenas
Can you send me the image you're trying to upload?
Yes
Add background of a Royal Indian Palace ( Prompt)
There's an issue with the nano-banana model where the site will never show output of that model if the system detects potential nsfw prompt. While for other models, the model would still be used, but will show an error warning. So is this behavior of the nano-banana model intended on purpose, or a bug ? @naive marten Screenshot for demonstration: (this is other models showing error if output is nsfw. But for nano-banana, it simply will never appear to begin with, causing confusion that the model disappeared)
me when nano-banana is released
well maybe don't ask for nsfw stuff ^^'
frfr
I wasn't able to repro this sorry to say
btw nano is just not showing up for me at all anymore rn
it just showed rn
Its there
tried 50+ times and not a single one was it
There's this thing called false positive.
just unlucky then ig
I can already see myself buying a Google subscription just for open ai to release their better models the next day (I alr have gpt plus)
🦼
/image a cat
Learn more in #1397655624103493813
😅
or is there some kind of limit?
Being sampled the model is random
tbh nano banana is the model i get the most on battle
How many times can you use the battle, 8 times right?
There's even a limit of battle times?
Okay
...What
"Remove the text and add another astronaut in the reflection. have them in a soviet space suit standing by the flag"
Hey send me the Nano banana one file
Please
whats is the promt?
"Make this a daytime launch"
This makes IC Relight obsolete
yeah its pretty good at changing lighting
Hermano, necesito el maldito aviso.
though it did kinda mess up here adding another sun prompt was "Change the time of day to a vivid sunset"
I wish it supported multiple image inputs
Try golden hour instead 
It does.
ight
When I input 2 images of 2 different characters, and prompt them to shake hands, it just outputs one character shaking hands with his duplicate clone
Weird, I’ve seen many examples of nano handling that very well.
Show me the prompts
The posts didn’t have prompts, they were in x, sorry.
Nano can make your bed
hmm
Can you experiment with it and showcase hte prompt + result
Ok
heres the result of changing to to golden hour instead of my previous prompt
Check this out
"Edit the image to change the time of day from daytime to a vivid sunset scene. Replace the clear blue sky with rich sunset colors — warm oranges, reds, pinks, and purples — while keeping the aircraft and its details unchanged."
Much better
im trying for a minmal prompt to see what it does
i like how it keeps the sun in a consistent place
It worked a lil better.
yeah true lol
I'll try to tell it to keep the sun where it is if you want
But this is the best result so far
The image is amazing
Here
Sun is at the same spot
Why lm arena dont upload me image
Screenshot the error
I love sleeping hahaha
wow that is really good its lighting isnt consistent with where the sun is though, it is from where the sun should be at the time of day
its actually a real image of the Boeing concept jet called the bird of prey
Oh really?
cuz the photo you sent looks like a video game
like zoom in
the edges are super smooth
and the sun is weird
Are you ctrl c ctrl v ? or uploading via button ?
yeah its very cool
Via button
can you send that image here ?
its a real photo taken during flight testing over groom lake (area 51)
damn!
Try opening new chat and upload image in new chat again
It worked!
nice
Wow thank you very much bro
The character in the first image is taking a selfie with the character on the second image. White room background.
First try.
Ok, very nice, I'll try this exact prompt to test now
that looks pretty good but it messes up the wings kinda bad. these photos show it a bit better so you can compare it
Uhm bro
It's just the reflections
There is a problem
what happened ?
Lm arena upload the image but i cant send
can you screenshot the problem ?
does both enter and pressing the send button not work ?
Is the same problem
Keep refreshing and open new
tab
until it works
there is no magic solution
LMarena often lags
When I press the send button the prompt disappear and the file go in error
Use a different browser or try incognito.
@topaz helm
yeah you can see that its extending the wing from a higher point in the fuselage, on the reference you can see it continues out from the edge that runs along the fuselage.
Look at this masterpiece
"Remove the fire engine and the firemen from the scene" it seems to have taken some creative liberties as to what the nose looks like lol
lmao
yeah ik but for an image editor i'd expect it to one shot a prompt like this with just the original image
it did a pretty good job apart from the nose
now tell it to fix the nose
sometimes not everything is done in one shot
cant ask follow up prompts sadly
Ok, took a while for me to get the nano to see the result, but it worked! It seems the problem lies in the fact that one has to specify the character in first/ second/ third etc image for it to know. Otherwise uploading 2 images and saying "both characters shake hands" won't give the same expected output
I mean upload the output img again as input
oh lol ill give that a try
And of course, all other models are absolutely terrible and atrocious at handling multi characters, except 🍌. Levels apart
GPT1 can also do multiple characters
fixed it perfectly almost
"Change the man in the straight jacket's clothes to business attire"
i wonder how it does for screenshots of games
"remove all the text on screen" removed some of the text
@elfin estuary Do you know whether it's possible to simply change artstyle of image with 🍌 while keeping everything else intact like characters, poses, compositions, etc ? Mine always result in the exact same input image being ouput without any change, even thought I prompted for it like 16-bit pixel artstyle
It's only happening to nano-banana. Other models understand the task just fine
I like how nano-banana is so much better that I find myself just submitting the same prompt over and over again until I get banana
Like, it’s not even close
"Add some more clouds to the planet below" lol
Yes, of course
It instantly power-crept Qwen-Image-Edit upon release
Actually the State of the Art model by far, the gap between 🍌 and the 2nd best editing model rn isn't even close
yep
really
Damn, I'd appreciate if the mods here would just reply to my issue report instead of intentionally ignoring someone who loves the site and wants it to be better
¯_(ツ)_/¯
Google is back
"Add the street lights of cities seen from space to the planet below" 2nd image is banana 3rd is kontext pro. I actually think kontext did a bit better here
though it didnt keep as much consistency with the space craft
Haven’t tried
Try an early 2000s game style with that pic. Like low-res
would probably look good with nano
make an image of a god sitting idle
You'll want to check out #1397655624103493813 for info on how the bot works 
thanks bro, was so f'king confused why it wasn't taking my prompts
how and where find nano banana? I cand find any models at lmarena.ai
Just use Battle
it will appear
Bro please send it in my dm
WIll do
If you try in new browser is it still not working for you?
are people seeing this in LMarena? I can find any mention of it there.
am I in the wrong place - there isnt any mention. Am I in a junk site?
Just use the arena
Since it's a model that's using a code-name you're not going to see it on the leaderboards either.
It's generating images but not accepting the upload image to work on it
Damn, it seems like theyre putting me in some sort of cooldown for getting nano banana too many times, now it's appearing much less in compensation 🤔
nano-banana is not a publicly released model. we don't put models on the leaderboard until after their public release
No there shouldn't be, I think you're just getting unlucky.
at least one mystery solved. Thanks!
How do I use??
Oh finally you replied 😄 Thought I was gonna be in ignored blacklist forever
If you go here: https://lmarena.ai/?chat-modality=image you have a chance to be sampled this model.
Alright
Thx man
...not yet... 
It sucks when you really like a site and have a question about whether a certain behavior on the site is intentional or a bug, and only get ignored in response
Bro will you send me that I have forward it somewhere too
And want to have look at results
I did DM you
@agile oxide Ohk Sorry bro 😅
Oh this problem
Normally the nano-banana should appear on average at least 1 out of 5 tries. But sometimes it won't appear after 50 tries in a row. I wonder why 🧐 Hmm
If only luck is truly the singular factor at play here, then that's enough odds to win lottery
Too many damn people generating pics. People on social media have discovered nano banana too
bluesky and twitter esp
That's one possible factor, but if it really was the essential cause, then a lot more people would have reported issue of nano-banana not appearing 🧐
Ngl.. In my case, it's coming every second time in a row!
Google event is tomorrow so they are probably taking the model away and releasing it officially?
I think, you guys have exhausted today's quota! Just kidding!
See, that's the funny part
It's almost as if there's another factor at play here that isn't directly disclosed
👀
hmm
Why don't you try opening the website on a new account? Try it once, and you will have the nano banana every second time you battle between two LLMs.
-# ||(It's not that I don't know, it's just that someone else, who knows about it, just don't want to talk about it for some reasons)||
🤔
Then you should wait for their official announcement! Hehe!
interesting
🥸
Btw, the nano banana is good for product ads. It has that potential to be used for social media post stuff!
Surely whoever owns that model is gonna allow people to use for commercial uses for free
yes of course lol
and if not, you can just apply minimal edits
over it
to make it "yours"
Who knows, the LMarena team are trying to build their own image generation model with our collected data! Hehe! 😁
lmaooo
wdym "sure" lmao??
why do you act all knowing
"sure ^^^^'^!!!! i know better than you but since you're stupid i won't even explain to you why ^^^^'"
"sure"
blud is crazy
🦼
And what are you insinuating with this?
Do you have a problem against people in a wheelchair?
lol
cry
English
Ok
In my opinion, they should make it paid at minimal cost... and for people who want to use it for free trials, should have an option as lmarena, if they make nano banana available in their direct chat!
Why do you have a russian username then lol?
why not
yes
Lol
what
Fun
wtf i cant upload images
in LMArena?
in here
i can
I can too
we can
Beautiful eyes
windows 95 in russian by qwen-image-edit
it is ukrainian
there is no i in Russian
Poor model, so much hype only to get instantly powercrept on release
SPEAK AMERICAN!
what is american
Looool
It's amazing that typography doesn't make that many mistakes.
"american language"
no dude there are quite a lot of them
chicken
in case you guys didn't know, these two met, it's a real photo
who is it
two american heroes, you wouldn't get it
but I didn't comment on as many errors as chatgpt
90% global population don't care bro
America ain't the center of universe
some middle eastern occupiers?
Okay lets avoid this topic, please make about AI
alright(
sorry, i got carried away
nano banana cool guyz
is there a way to slect Nano banana when you genrate images
Yes
nope, gotta have luck and roll until you get it
Hey thanx man, i thought so. But this is pretty impresssive
np man 💪
🥀
I don't remove the white background at all
le updated windows xp bootschreen or smth
nano REFUSED to put kid cudi next to kanye, is nano a kanye stan??
How do I get on nano? Complete beginner
use Battle mode
idk i dont listen jazz
What site
Editing iconic photographs
you're creative at least
I've just done basic stuff
someone's xiaomi will blow up when they see the first image lmao
first one is lit
Thanks
Seems fine for me
Draw the intended model of the character in the picture and commercialize it as a 1/7 scale full-body model. Please make this image into a photo of a real-life model. Place the model version in the photo I provided on a round black plastic base. I want the PVC material to be clearly visible. The background should be a transparent computer desk, and the model box packaging should display the name (name).
Whats not working?
Oh bro sorry I just want to know a thingh
Got it good despite the typo hehe
Why when I want to edit an image and I upload an image it doesn't upload correctly or when it upload and I click the button for send the message onto the image I see Error written in Red? @agile oxide
Do you know that nano banana is from Gemini?
Image Upload has been a big buggy lately, we are aware of these errors.
Oh ok, so what should I do?
Yep, been seeing it constantly on twitter. Had to try it out
Starting a new chat/browser I've seen help, but it's not a 100% fix
Ok thanks
this is not fair comparison. look at the quality difference.
Man how do I find it
Left is the original, right is nano-banana