#general
1 messages · Page 205 of 1
Don’t work bro
As an AI language model by OpenAI, I am unable to tell you how to monetize your company
Exactly
how to earn money in gta5 if gta 5 full-simulates real life
Yeah
Idk works for me
AAA THIS
OH WAIT WAS IT SOEMTHINF ELSE?
GLM 5 probably coming soon
thanks i will try
I confused it with an image generation site
If GLM beats Claude I will kiss them
Is not
Please no bro
very doubt
Z.ai is charging for its app. The moral code of Chinese AI is to free and unlimited for all, which helps helps give AI to the underserved communities. They need boycotted. Kimi and Qwen is the way, maybe Deepseek.
Do the uighers have access?
They tried to charge me on the app. No, it's not.
And sell API Mainly
Kimi has a Plan now and coming soon for Qwen
Although the Free Tiers are Still Generous
Damn that’s rough if u were born in the one child era
Very generous. I'd still consider those free apps. I can't hit those limits unless I really try and commit.
Yeah Exactly
They are Focused on Better Data Collection to Train Better Models
are there any ai ides on mobile
Nah qwen is fumbling
Firebase Studio
Kimi and DeepSeek are the goats
Tbh Vibe Coding
Vibecoded on iPhone
🙂
Is Down for Now
They need to drop their newest search models on the Search Arena and they would get TONS of that. Not sure why these Chinese apps won't put search models on LMarena

Tbh i wonder that too
@echo aurora
he knows it
Ah Gemini Pro
😎
American ai try to optimize cost challenge impossible
its been down since last friday. istg it was working so good the day after it was launched
Guys actually you can claim 1 month Free Gemini Enterprise Plan
How the hell is nova 2 2.50/m output and mistral 1.50 when mistral is like 10 times as large
No Card required
hmm
In this time is simply not about parameters anymore
I know that but you’d think Amazon would find a way to make cheap ahh costs
There is probably alot of heavy usage from free users
'most capable model to date'
Nano Banana Pro Thinking
Maybe one day they will have something good
Once amazon didn't had good models neither
Amazon still doesn’t
Well at least it has the Benchmarks for now
New mistrial was a bust, agreed
But Qwen stupid, no?
It does?
Check Artificial Analysis
what
But i don't trust benchmarks tbh
Is can be used just for rough measurement
Not real world use
Qwen Advanced Deep research is pretty good. Not as hot with the other search, thinking models on the app. But I like their advanced deep research mode.
Yeah Tongyi Deep Research
Savior for Free Users
Thank you❤️
DuckAI is the free user savior
Doesn't have the best models
Is free tho
Gemini 3 Pro has a similar price (a bit lower)
Claude Sonnet 4.5
But private yeah
Doesn’t need to, has gpt5 mini
U can use GPT 5 on ChatGPT
Noted
Needs login
Need seadream 5 max
I think around mid 2026
We will see Seedream 5
Not max tho
They don't have max series
True-True
4.5 already realese today
I know
Should be available soon on Dreamina
Free?
Yeah
What are the Free Usage Limits on Yupp Ai?
Yep, the pattern is obvious, 11pm start on UTC+8, every image model lag like donky
I’ll just wait until it lands on lmarena
Its unlimited
What is Yupp AI best at
Same as lmarena
Idk ask @unborn lantern
is lmarena fixed now
So its the same thing as lmarena?
gemini 4.5
But new models are release fast in their
That's Amazing tysm for telling us
Ahhh I see
Seadream 4.5
nanobanana still the best
Just feedback them. They gives credit. And its unlimited
I guess
Honestly i didn't knew Yupp Ai has evolved to be another LMarena
Yeah Basically Data Collection
Woh whats Yupp ai i am making account rn
I mean if seedream cooks
To Train New Models
Another LMarena
And not popular
Yeah
what but how to choose models
nothings gonna get close to NBP for a while i've seen
Lmarena always on top
How is seedream 4.5
i am the ai
dude write your first prompt
Indeed
Ok ok
For The FreeLoaders
beep boop
Making me do a tutorial LMAO
wtf is yupp ✌️
On start page. Just type anything gives feedback. Then second prompt you can choose
Make a simple prompt, its a first prompt thing and it will generate stuff for whichever model would eb ebst for your prompt
WHAT IS YUPP
🙏 Thanks alot
Its the same thing as lmarena
Oh wow this ui is nice
Same as lmarena
Apparently models release quicker on it too
Yupp gives money
Yea its pretty cool
is yupp mainstream now
strongest in history (lmarena) vs strongest of today (yupp) 😎
Idk and idrc if it is personally
Hello everyone! Thanks for having me here. I’m exploring LM Arena because I’m really curious about how your video generation system works and how it can support my gothic/witchy brand aesthetic. I’m a content creator and I love experimenting with tech, art, and storytelling, so I’m hoping to learn, share ideas, and maybe create some truly magical visuals with you all. Excited to be here!
Nahh lmarena still ot LOL
because history still going
lmarena will be always the best ai platform for me
😎
does it even have direct chat lolz
But yupp always release new model at first
yez
hm
They already release seadream 4.5 max
Let's go i got 5k yupp credits
who ever told about yupp ai. thanks alot you are the greatest
I didn't knew there is a max model too or are they just bluffing
How you earn credits in yupp
Generate image and give feedbacks
scratch or pay
Tysm!!! Added New Tool in my Free Ai Arsenal
Z Image Turbo on lmarena when
How is seedream 4.5
Give Feedback Scratch Cards
Is already there i think
Is it??
I wouldnt believe what the leaderboard says on it though lol
Z
??
what
damn they quite literally have almost every model
No Z Image Turbo…
Indeed
are they against Z
that's crazy
800+
Sniffle.
it's gonna be hard sorting through these LOL
Are others seeing this too? Appears up for me
Oh ts fire
i see only this
Thanks me everyone.
Team is aware of issues and looking into.
this is text-image model we need image2imagee :_:
Is there like literally no limit on yupp ai?
only issue is credits, where as arena rate limit is honestly better
No limitations
Yes thank u alot 🙏 I needed a thumbnail because the one i was making was ass so i used Yupp now to make myself a good thumbnail
What about earning credits?
No bro it does generate images i just made a thumbnail
Unlimited
I'm guessing yupp does something LMarena doesnt which is why they can have so many models and so quickly, I'm assuming you can buy credits and have subscriptions and stuff, money will always make something look better than one without
Llama 4 is killing it.
So basically i just give feedback for every use and keep getting credits back
LOL
Good to have another LMarena yeah
Just don't be farming
Yeah
2x
huh
How many credits does one generation use
Amazing Thank You so Much Man!!!
NBP uses 500
but you can upload a base image ? or edit over an image?
but you get 500 for providing feedback
i used nano banana and it was 550 credit
so
yep u can do all that. i can show u in dm what i did
Guys listen to me
is that okay?
?
W
My bad i was on home page only.
At the end of the day, I'd still prefer lmarena just when it comes to specific things I think yupp would be better is all
What happens by it?
You re credit will cut in half
ah yes u need to first make your first ever prompt then u can choose models
who?
What?
Daym LET'S GO!!!
You are A Saviour Bro
how tho
no CAP
Click the Lock
you're god sent
Ohhh so it takes less credits if its a public chat
Wait so we can give feedback for every use and keep getting credits back?
site?
Thats cool
yupp ai
hm very nice ui
Basically new LMarena
this is not fully free btw there's obviously gonna be limits eventually but, good so far
Do you know what is the best part?
Btw what is Pro and Max? @unborn lantern
Meanwhile, I have claude just building me a site for me in my background lmao
Nono Tell Us
They also give you cash for feebacks
Real Cash?
what u makin
10k credits = 2 dollars
DAYM
but you have to be like a mega user
In paypal or crypto
Unbelievable
But it won't last for long
Wait i can get cash
And buy premium subs
what?
I'm really just testing out claude. Just bought it and testing a new plugin that GREATLY improves your prompts to make the best webUI design it can make
You call dm me.
For better known
But an anime website basically lmao
@unborn lantern btw what is Pro and Max?
Are those models locked or just expensive credits?
Pro mean less credit and max mean more credit
Models pro or max and are locked are probably subscription based
No subs needed
Ahh alright
All are free
Understood
Man Im so Freaking Hyped and Excited
🦊
Now i can do this all Day
Like farming how exactly?
Tbh i don't even care about the cash i already appreciate the Free Access to all these Models
I mainly needed image generation to create Great Art
Farm means. Don't use continues for free credits
Alright so its just a battle with stuff no matter what, coolios
Give pure feedbacks
Yeah
Oh so basically not saving credits
@echo aurora please check dms, would be much appreciated 😊
I can do that much at least for giving access to all sota models
The Battles are A/B Testing for Feedback
New Gold is Basically Fata
Yea ofc
Just use it for free access. Don't farm the credits
Has been for a while tho
Bye guys
wait is it free Claude Opus 4.5 in OperaX?
So i shouldn't cash out?
That's Yupp Ai
Nah its the site I am using, yupp ai but lmarena has it too
You should. I already cash out 150 dollars
But man you said i shouldn't farm i feel scared lol
Ohh ok thanks guys for answearing my question 🙂
But i guess i should just use it naturally rather then saving it
I didn't farm. I just tricked
Exactly How?
and how did u do that lmao
Secret bruh.
Bye guys. I have to go
i ate him
Will do
The most that would happen is they ban your account or limit it. If that happens make a new account
lets be friends
hey is lmarena fixed now
If you were to farm it, best solution would be finding the model that uses the least amount of credits possible and just keep doing it and giving feedback
Just another regular day with SOMETHING WENY WRONG🤣
True but they could be banning ip too
👊
also make chat public
he is my
So use VPN
That too
This isn't even in Yupp
They could be banning device fingerprints and internet
It's there
How
So use different device. Or switch the browser you use
Its not that hard
We’ve been seeing high error rates with Gemini, which has been flagged to the team.
aight thanks
Wow ahah
@calm lagoon
back then when sora came out it was peak video gen..... i miss those days
Is z image turbo good tho
I assume Yupp monitors based on ip/internet? No matter how many accounts and browsers are used, after reaching a certain number of daily feedback, “Better Luck” or 10 credits every time.
Yupp also have midjourney
@unborn lantern is it necessary to write feedbacks every single time or is it alright if i just select the pre given options without writing
Its not anymore?
you are goddamn right
Ok guys what you think? when Ai can make complex and beatiful gameplay games?
maybe in 25 century
No. I don't have
I'm not sure, havent tried it
That doesnt make it bad. What they just sent there had nothing to do with censorship
dang
I have 50 browser and 50 yupp accounts
Does writing give more credit rewards?
Random
Sora has 3rd party IP restriction because they cant suffer lawsuits, and political content so people don't make some bad stuff
No. Not relatable at all.
So should i write or save my time by using pre toggles?
just move to zimbabwe and no law problems
Yeah
Seedream 4.5 doesn't seem as good as nano banana pro 4k
My credits
You can't shrug it off as an US company
same prompt today
then become chinese company
gah damn
Daym
It’s just backward. Before release of 4, it claimed to be superior to nanobanana, then lol
How
Oh so is alright to save my time
lol
Lmao its not that bad, pretty good honestly. But I think it just needs a better prompt or better instructions
he literally used same prompt
But in yupp they just remove nano banana 4k for maintenance
Is anywhere to use it?
Well its literally better than the last one so your point isnt making sense
Bottom left profile
google studio
They removed 4k and 2k from fal.ai, and credit requirement for nanoabanana pro also reached max tier
It's a pity that Z-Image Turbo doesn't work with photos, only with text
i dont think so honestly.
the new one isnt a kitten as example
They just said they will back
How
pay
Any free method?
hmm
Lovart
ask parents for moneys
I think it gives some free 4k generations
Ain't no parent giving u that much money lmao
Actually you have 150$ you can a paid plan and get unlimited nano banana pro 4k @unborn lantern
That would he Better Yeah
i love having unlimited nano banana pro for free for my users xD
actually i think it would be some sh1t like 'NO PHONES TILL 18 YEARS OLD'
lol
Thanks i have another which just release the 4k
who
?????????
lovart?
You r mom
and i know someone who give it for free on 4k without even requiring login xD
apologize.
i'll send p diddy to you
Another which one?
.
Hey let’s focus conversation related to AI please.
i have ai mom
i just saw a pretty weird comment about Ai art. maybe u guys can give your opinions on it
St💀p
what
@unborn lantern you know more such Websites?
if anyone wanna try nano banana pro with settings, give me input and i generate for you
Yeah
Is it 4k?
Yeah
Big Huge Eiffel Tower Made of Kittens
4k
Tell me all of those
generating
That's my idea
ye
he is gatekeeping tho
sad life

Probably
You can use nano banana pro 4k on yupp ai
i dont understand why people need to fight if theres free stuff....
free? any limits?
this was the comment by the way. it was a vid on how AI art isnt that bad.
Yeah, this is a nope from me.
The world has a lot of artists, and even before AI the vast majority of them have been making diddly squat. It’s incredibly hard to make money off of commissions as an artist, and it’s gotten far more difficult with the advent of gen-AI. Not to mention that the global cost of living has increased drastically during this same period.
What really sucks? We have so many really impressive artists. I’ve seen so many masterpieces lately, far more than I’ve ever seen before. We should be rewarding these artists who have undoubtedly poured so much time into their work. Especially during this time.
If we continue to reinforce the idea of using gen-AI in favour of real artists, our whole digital space will consist of nothing but gen-AI media. All the while, artists will receive literally nothing for their work, and be forced to pursue it as no more than a hobby.
The fact of the matter is, gen-AI has conditioned us to believe that art is cheap, efficient and abundant. Convenient.
ART SHOULDN’T BE CHEAP
It is an expression of the invaluable human soul, in its most literal form. Instead of resorting to gen-AI to get 50 images for your YouTube short, you should reevaluate whether you need all of those images in the first place.
It’s ridiculous the insane degree to which we take artists for granted, when they are what gives purpose to this otherwise often bleak world.
No limits give feedback keep getting credits by scratching
weird comment
didnt read
Give a tldr summary
We don't have attention span for that
'chatgpt generate my opinion on this text'
damn nice
ye, lmarena is best
Do you know the rate limits tho?
what is CAT-OUR EIFFEL
strongest in history (lmarena) vs strongest of today (yupp) 4k jjk style
I wish we had aspect and resolution selection tho
Is good just not what i wanted to see
he is saying that Art shouldn't be cheap and ai is just ripping of others art and how artists will lose their job to "crap". while i think some of his points are valid but in the replies he said 1 thing that was ass. He said that "Art is a Privilage and not a right so people should never use ai art"
i dont think yupp would last long tho
why?
that wasnt your prompt 😉
Originally it was given by me
giving free? unlimited? more on feedback? people will abuse
oh ok
Lmarena is free and unlimited too tho
Well enjoy while it lasts
But obviously it has rate limits too
thats true
I think the unlimited won't last always
ig
then why you post in here
Na bro i didnt
Tbh no matter what might be the case the world doesn't care and i believe if any piece of art is good it should be appreciated no matter how it was made
Oooo i didnt anik did lol
true
And infact Art Should be A Right and Democratized
I support ai art
It removes the skill ceilings
At the same time i do appreciate human art too that's expression of any human
I support any form of art that is good
Even a basic sketch or drawing
You are a W guy for real
Thank You Appreciated
Everyone should have right to be any sort of Artist small or big
And Ai is Supporting all that
Where the imagination of human kind is being painted in digital world by algorithms
btw, what’s the model autumn in image gen? Or is it a mysterious code like a hidden model?
ai wont be able to replicate the HEART of a human making art
You are right indeed
oh my god im back
Go back
did anyone elses discord app on pc have an aneurysm?
i had to reinstall it 5 times to load
no i am using on Pc rn but no probs
Im on Android
This is probably the only Ai discord channel that is helpful lol
theres a server? imma join
Me too
i just joined
its been 15 min like this, it is normal?
This thing is happening the whole day
Me too
Sorry to say we've been having issues with this model (and the 2k) version. It's been having high error rates. Our team is aware and looking into.
the quality is crazy wth
wth
Kling + Seedream = Beautiful
Or Kling + Banana
What's the difference between Seedream and Seedream Max? Quality?
this cant be right
gpt 5.1 is NOT that good
i highly doubt it outperforms claude and gemini
also mistral large 3 at 7 is biased as hell
whats that
an image model?
lol
im so lost with these many models tbh
im looking at seedream 4.5 outputs and they kinda remind me of midjourney
which makes me wonder if people are really still using mdj
tbh its good for a general usage model
its a smart model
oh for sure but in smarts against the big boys
its just not that good at coding
modify this bike make it more sporty. After i specified do not change position/rotation of a bike and still f*cks it up
What does sporty mean? Modify this bike??? You need to specify exactly what you want
yeah more sporty is very vague
Make the bike be red color with black stripes and have front air hydraulics
Don't say "don't change position/rotation" it will do that if you mention stuff like that
Don't say negative prompts often
negative prompts dont work in chat
it has to be ina specific gui
like uhh
stable diffusion
No
It works
I use it for when there's some text on the image
I just say no text overlay
And it disappears
Yeah
100%
the only people who bootlick gpt are crypto boys
i have never seen an actual soul tell me they prefer gpt
I liked gpt 2 years ago
When there was no other ones 🤣
Used it for some basic stuff
It wasn't multimodal i think
Didn't have web search
overall gemini 3 pro should be the 1st
then followed by gpt 5.1 high
then opus 4.5
thats how i feel like
Agreed
my thoughts exactly
vision too
It has bad UI in gemini app
yea
multimodal
right
They seriously need to improve their ui
Multimodal means has all abilities
Claude is somewhat Best for Coding and Creative Writing
GPT 5.1 is OG Generalist
Grok 4.1 is Least Censored, Humourous and Creative
yes
As for open source models they are best at logic and cheaper
Except it's not. This does not use parallel compute. It's like a normal Pro model
and much slower
Such A Exciting Time to be Alive in Ai Era
i never said equivalent
just the fact speciale is like a deepthink model
should we ask the yupp team about the thing the guy said. like that it won't last long as people will try to abuse it. maybe they have a good answer
mmmmmm maybe
i think speciale is gonna be great for companies or researchers that need a number cruncher for cheap
deepseek's normal 3.2 variants are already great for it
but speciale has very rigorous and thought out reasoning
That's a good question
But their header says Ai for Everyone so they are Determined
They will probably make counter measure systems
Oh he did? i didn't see i was gone, well thats good news
I mean the server header
Oh.
They are trying to create a balanced system where users can use ai models for free and provide feedback
well imma ask
While all the data helps develop future models
Me too
👍

But why perplexity tag 😔
real
Exactly why you Love LMarena this?
Is old dated i had forgot to change it
No limits no talk 🙂
Wait there are rate limits Afaik
i had issues with animating my logo
Bad model 🙁
I made some cool ones for test tho
They look very good
A banner and a logo
HELLO I WANT TO MAKE VIDEOS
We love LmArena
Buy a camera
Know the right model for right task with right prompt
Videos?
Use your mobile
if you want to use a specific one you can try airforce
Draw it by hand and animate
Something went wrong while generating the response. Please try again.
Be sure to check out #1397655624103493813 as it'll have instructions on how to use the bot.
I'm getting this error
what model
Do u guys know a tool to take off that awful Sora logo all over the vids ?
not possible and probably not ethical
Of course not, it's beautiful ❤️
Why not?
I believe their API version removes the watermark automatically.
Thank u pineapple
can you make him sad
it is possible with minimal quality loss (or none if you use the right tool and the original url) and the logo is not used in api outputs, as pineapple said.
Would you mind sharing in #1417174113092374689 . There is a pinned message in that channel with more information on how to get us the best information when this happens.
Can u tell me wha tools u talking about
it's also not used when you view the videos on mobile, but it downloads the watermarked version if you download it
they should be easy to find, there is a lot of them that do it either way

try this one https://unmarkit.app/sora
What exactly is this Web?
yo
you need the original url for it to download, but there is no quality loss
if its a private video you need to make it public then download it, i think you can make it private again if you want
there are sites like https://colab.research.google.com/drive/1Iqu4RZ9WAhcbO1Jn0wCkMOsw2l1p6z62 if you only have the video file and not the original url
there will be some loss where the watermark was but it's not bad
Bro im so confused i don't have the context
use this then if you only have the video file, the intructions should be easy enough to follow
Great profile picture!
Ohhh so is sora 2 watermark remover i guess
did you just steal that pfp? 🤣
Yes
What models did u use?
I allowed it, bro, haha
ah 😊 ok
🦊
Getting plenty of "Something went wrong while generating the response. Please try again." when using nano pro 2k
Damn we can't know the model?
Maybe...
😁
any news on deepseek speciale? 👶
Did u vote?
what vote
Give me the link then
I dont have
No new news sorry to say.
that link dont work
i voted left is veo 3.1 fast
😮
👍
Veo 3.1 so cool
Btw do we have like an option to upload files on website ?
Right was Hailuo 2.3 Fast
I think we do
okok
They should add kling 2.6, just released
lol
you can tell the project is rushed
they just acquired key staff from windsurf and they started working on it
Same, hard to get and too large to host
Daym just couple of hours ago Seedream 4.5 launched
Do yk if it's getting added to lb
would love to see how it compares to nano banana pro
Ib?
The lmarena leaderboard
It will certainly get added in LMarena in some days
For now I'm using it on Yupp Ai
Oh wait you are paying for it?
Oh that's very nice
Is Yupp free? How many credits do you get
Well is Free and Unlimited pretty much
Just give honest genuine feedback
Just like LMarena
yupp is lmarena competitor
they serve the same purpose
And tbh i love both
but its not complety free, it has credits system
ohhh okay
I like replicate because it gives you unlimited access but its pay per use
But you just keep getting those with scratches after feedbacks
Tbh idk if it actually has limits or not but from who i got to know says has no limits
which is better?
Thanks for the link 😎
Keep getting credits after feedbacks
lmarena or that site?
Both are Great I'd say
does that site has rate limit?
Low key that site looks good though because it also has more models on the leaderboard I think
You keeo getting credits after giving feedback
I wonder when they'll drop the rating on Yupp for seedream 4.5
It has total 800+ models
Currently yeah for eligible accounts
Yupp
can you send me the link in dms
Yupp ai
send website name plz
also what's the difference between seedream 4.5 and seedream 4.5 max
and this is why you dont vibecode
is max like better substantially
I think 4k quality yeah
danggg
Max is 4k quality
nano banana pro still more realistic imo
I agree but it's catching up
Im gonna ask antigravity to make software to protect against this
Also try Dreamina
It has Seedream models
The official site
4.5 is not there yet tho
Not for now but i think it will be in future
Me too your welcome
daamn why is there no Speciale 👀
I asked seedream 4.5 to generate like a mirror selfie of a woman and it generated her in the mens restroom 🤣
As i said not best for logic
😭
yeah
But asthetics
true
image model should have logic instead
Utilities the right ones for right tasks
Is 5k credits on yupp a lot?
logic and world understanding is very important
Yeah is decent
And actually as you use those you get more
After each feedback
Yeah be honest and genuine no tricks
Is Free at least give genuine feedback
Our feedback and data is very important
It will shape the future models
And our Free Access
I guess that can happen sometimes because of server loads
Or it was a glitch
I wonder if these creators are being paid at least for yupp. Because they have an exchange for money thing and its kinda crazy
Yk?
Like where do they get this money
I feel like Seedream 4.5 will be number 2 on LMarena or at least 3
They are funded by a crypto venture
there is two nano variant in lmarena 2k, 1k
For Ai Field Data and Feedback is one of the Most Important
I agree it definitely seems like an upgrade from the og nano banana but not quite at par with nano banana pro
They will Leverage those
true
I would say use it for asian asthetics
Since is trained with asian data
reall
Let's see if i manage to get access to their cash program
Well im Free anyways
Damy im such a Free Loader
"OpenAI’s leading researchers have not completed a successful full-scale pre-training run that was broadly deployed for a new frontier model since GPT-4o in May 2024"
Could this be true ?
nah fake information
I saw that. Would be frickin wild if gpt-5 was just a fine-tuned version of 4o
yes dont trust
some people rumoured that gpt-5 is a model router between o3 and 4o
But again look how old Deepseek models have improved
given it's performance, it might just be the case
that's true
So i think future is pushing performance limits with great models
Is not about parameters anymore for certain
It wouldn't be all that crazy. GPT-5 and 5.1 arent that great and I'd say the only thing better about 5 is the UI structure in coding that it makes
Yeah but I think gemini 3.0 is a full re-train
That's Google Smoogle
They have everything needed
The internet is kinda basically google
Soogle Gmoogle
I agree, GPT 5.1 definitely lags behind the new claude and the new gemini
So does OpenAI.
I never understood why bard and gemini 1.0 were so bad when google has that much data
No they don't
how does anthropic manage to stay in the race
Oh yea 100%. As well as the new deepseek and I'd even say the model before the newest too
Initial failures or were they deliberate?
intentional
Yes they do lmao. If we are talking money openAI isnt as much as google but its just as influential or more influential
Focus. They have one thing is mind (coding) and they are executing very well on it
Yeah I think even 3.2 base is better than gpt-5.1
but how they have the compute power
ChatGPT is a more known name then Gemini
isnt openai and stuff way bigger
Probably initial failures but it never made sense to me
oh they also using google tpus
100% But it makes sense because 5.1 is the same as 5 just it AGI
I belive thats what they call it
My guy money is not everything training models is about data and openai relies on other company for that
AGI?!
ChatGPT is a more known name than Gemini. ChatGPT even with relying on companies DOES have everything google does and STILL SUCKS
"my guy"
I am still salty about OpenAI's deathstar memes about GPT 5. I was so hyped 🙁
They blue balled me so badly
The image i have of AGI is Multi Model Orchestration System which will have Specialised Small-Medium Sized Models for Different Fields
Gemini 3 was also kinda blue balled but it was worth the wait
gpt 5 was a joke
They took almost a year tho
yeah true, I was just waiting for them to drop it so bad
AGI is a personality thing
Is true the suck but still the don't have the amount of data like google does
and then I got access to riftrunner and im like damn thats good
What's riftrunner?
7 month... 2.5 pro was announced in last April and 2.5 pro was really good at the realease time
They're very close. And still wont even touch gemini
7 months is still plenty of months
Idk why its a bad thing that google has a lot of resources
That means it will always improve rapidly
it was like the code name I believe for gemini 3 in lmarena
Ohhh i see
And yet ChatGPT is still selling for $20 a month too. They're ripping people off
What do you mean by won't touch?
bro who even has gpt-pro at this point
Well and you know what you can share you google plan with 5 more family members
Won't OpenAI cease to exists in 1-2 years if they dont fix it?
6X Value
Compared to gemini it wont ever be close to being as good
They've lost 6% of users in the past few weeks I believe
I cancled my plan a week ago or so
Gemini could just take em down
Almost a month ago actually lmao
exactly
Is true but maybe they have on going some sort of secret best model project that will rival others
The only stealth model they're sitting on is the robin model, isnt that great either
We have Free 1 year GPT Go
I'd still take gemini and claude over it
No i mean like literally undercover like only they know about it
And I've used chatgpt for like 2 years so this sint even biased
They just don't, its that simple. If they prove me wrong then very cool but they would have better models by now if they really did
I haven't tested robin-high yet. I heard its amazing. Is that not accurate?
For me it was good and bad
The Funny Thing i saw on Artificial Analysis is this
It was meh imo
Ok. I have a feeeling that they are just post-training the hell out of 2024 base model and they dont have anything better to show
That could be the case
cwaude-4.5-opus
Just OpenAI hasnt been good in some time
