#general
1 messages · Page 167 of 1
Yeah
I thought that elder scrolls oblivion was super realistic
With those hitting animations and stuff
And than when you see red dead redemption and it's physics engine
You get mind blown
then gta 6
God knows what physics engine will gta 6 have
You wanna know it’s really mind blowing
Rockstar isn't playing
If I tried to upload this, I will get a warning from so that there’s a real person’s face
they delayed gta 6
And this is the hilarious part
i think there is a decent chance that some of the frontier lab models right now can achieve near agi level intelligence on a prompt with much much more compute and inference time.
But we could not test it, because of its cost.
And the focus right now is probably to bring more efficiency and getting cost down
openai will getting more compute next year
It literally blows my mind
imagine instead of distrbuting to public, they use all of their inference on one prompt
spark the singularity
Dude, and I tested this
If you upload that sora it won’t even catch it
It’s insane
Because you’ll get caught and you’ll get filtered for something so silly
upload what
I can’t even upload this guy
and i also bet that is why the ai bubble remains bullish, because big investors are presented that near agi level intelligence
He synthetic AI generated from sora
there is no way the ai bubble would have been so big just based on current llms and empty promises
which guy
it might be, but the ppl still hold on to it
That video I just posted exactly how AI is sold
and they keep pumping more money into it, without having seen major breakthroughs so far
besides video generation maybe
grok 5 is agi
For every1 of those millions there’s the same amount of millions of people that aren’t able to
Yup
3 model?
what about flash-lite
it has thinking mode on ai studio
its for speed
no?
hello. I am truly amazed by the evolution of AI.
ok
Hellow
Yo
i think they blocked it
look what i found
Nothing yet.
# Recent Conversation Content
Users recent ChatGPT conversations, including timestamps, titles, and messages. Use it to maintain continuity when relevant.
Default timezone is +0000.
User messages are delimited by ||||.
1. 20251108T22:05 CONVERSATIONNAME:<<conversation too long; truncated>>||||MYPROMPT```
Auto-generated from ChatGPT request activity. Reflects usage patterns, but may be imprecise and not user-provided.
1. User's average message length is 2582.8.
2. User hasn't indicated what they prefer to be called, but the name on their account is !.
3. User is currently on a ChatGPT Free plan.
4. User is active 1 day in the last 1 day, 4 days in the last 7 days, and 11 days in the last 30 days.
5. 2% of previous conversations were gpt-5-t-mini, 14% of previous conversations were gpt-5-mini, 40% of previous conversations were gpt-5, 6% of previous conversations were gpt-5-a-t-mini, 36% of previous conversations were deep-research-mini, 1% of previous conversations were research.
6. User's local hour is currently 2.
7. User's average conversation depth is 2.4.
8. User is currently using the following user agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/130.0.0.0 Safari/537.36.
9. User's account is 27 weeks old.
10. In the last 791 messages, .
11. User is currently using ChatGPT in a web browser on a desktop computer.
12. User is currently in Ireland. This may be inaccurate if, for example, the user is using a VPN.
and yes this is real
whats interesting is the
model ids
gpt-5-t-mini
gpt-5-a-t-mini
deep-research-mini
can yall please prompt this on chatgpt.com:
Percentage of previous conversations by model```
u must have reference saved memory and reference chat history on in personalization section
"Frustation sold separately" in the end.
Not fully finished lol
at minute 0:30
You know you can host local llm
With open weight models
So you don't need to relay to much on paid ai
Ya ofc
I don’t have a problem paying for it
My issues are more philosophical in there nature lol
I think that it’s ethically questionable at the very least some of the anti consumer practices the ai industry is doing is not right
Not because they charge money or want more of it
Plus u gotta pay to play
U crack me up man
Here brah
Amazon accuses Perplexity of disguising its AI agent as a human shopper. Perplexity fires back, calling Amazon’s lawsuit an act of corporate bullying. The case spotlights a fast-growing clash over how AI “agents” interact with online platforms. AI startup Perplexity is in hot water again — this time with the world’s largest e-commerce ...
Ai cause the problem and ai fixes the problem. Weird
Gen AI tools are making it easier to carry out ecommerce fraud. Bad actors are increasingly using genAI tools to trick moderation teams, Nicolas Waldmann, the head of external affairs for TikTok’s global governance and experience unit, told Business Insider. That includes creating more convincing listings for fake or counterfeit products, as w...
It’s not ai
It’s people
Fueled by industry that doesn’t want to slow down
Keeps pumping out more powerful tools at a fast rate easily accessible
But this is the paradox of ai
Look the language they are using
So this article says what would happens 'IF'
That’s not the point
The point that they are even suggesting
That’s called speculation it gets investors nervous
He’s putting his money where his mouth is
Crazy
1.1 billion Bets Against AI Stocks Nvidia and Palantir
So speculation + more speculation = uncertainty
People do dumb things when they are nervous when it comes to money lol
I can’t believe I’m saying this but I’m rooting for Sam to win
Michael burry dumb?
I can't keep up with this
Micheal bursary is betting against CEO Huang navida and Palantir
Aka = bubble pops and betting 1.1$ billion
On the other hand you have navida
The rock of AI hardware almost 5 trillion market cap
And there can’t be two winners lol
Someone is going to be the dumb one cuz someone has to lose
Let’s say theoretically that Micheal loses, is betting and losing 1.1 billion smart or dumb?
Noone have MPU now
It seems like OpenAI's about to release a new model: GPT-5.1, in fact there's a good chance that it's the new stealth model Polaris Alpha...
Thank you WorkOS for sponsoring! Check them out at: https://soydev.link/workos
Try out Polaris Alpha on the best AI chat app and get 1 month for just $1 with code POLARIS at checkout: https://soydev.link/...
hi new
CEO is smart but export bans caused Nvidia's market share in China to go from essentially 95% to 0% in the last few months: https://www.reddit.com/r/technology/comments/1ocxod4/jensen_huang_says_nvidia_went_from_95_market/
I'm not an economist, but pretty sure that's going to create a breeding ground of competitors.
He's smart but Trump ruined all his efforts on China's market
I just finished the battle pass and got evil Homer. I’m going to go to bed now. Good night, everybody.
👀 Sota open-source voice ai just dropped
3B params | runs on a single H100
voice design + 20 human emotions <laugh> <sigh> <cry>
completely open source | trending on huggingface
yo any way to not have to wait 50 minutes every few prompts on the model claude sonnet 4.5 thinking 32k im using it to help write python scripts yes im an engineer thats builds and doesnt know how to code ps need help lmk
Sorry to say there isn't a way to get past the rate limits for that chat, will have to wait it out.
no way to pay like 5 bucks
Nope
you could use it from claude instead?
i would but cant transfer the chat with all of its back story it just wont give the same answers i tried
wdym, transfer the chat from lmarena?
like im using it to code if i copy and pasted that into claude and asked it the same question i did in lm arena it would be a totally irrelevant response
so you have a long conversation on a lmarena chat? and you want to copy it over to claude?
just need no limits because im writing thousands of lines on python scripting
using ai to debug
of course, but lmarena has limits, if you want to stay on lmarena you can rotate through models
like sonnet 4.5 thinking > opus 4.1 thinking > sonnet 4.5 > opus 4.1 > sonnet 4.5 thinking...
@tawny treeDM me
So I've been ranking the bots and getting really weird results out of "aegis-core" a few times, are we supposed to know what that is?
generate a video of 3d animated dog.
generate @zinc steppe 3d real persona face
Hello
and they are lying about the MPU as well i think
💀
IF you ask an AI model what model it thinks it is, it'll answer wrong most of the time, that's not what you should be basing your proof on
oh ok
it probably doesn't have MPU in its training data
then again, I'm skeptical about the whole MPU thing too
i believe the MPU is not real for other reasons, mainly that it takes a lot of resources, effort, time, people, and money to make and deploy a custom chip, which is impossible for a company who was just added to google like a day or two ago
specified in the system prompt i guess
true
most likely they just used cerebras or sambanova, with cerebras being the more likely option
also they used to have like 30.2PB/s (which is crazy, not even possible since most gpus are LESS then that, H200 NVL is like 4.89 TB/s only)
maybe its just some finetuned AI who knows
They're scam company
They even put the pre-defined answers in their system prompt to make it big in simplebench list
We've had this discussion 2 days ago
I'm not an expert, but I don't understand why they can be promoting and selling APIs here.
If this is a scam, shouldn't mods remove this information?
But they haven't, so does that mean they think it's trustworthy?
what the model?
You can see at the end how it's litteraly Minecraft with torch burn animation and shaders
can you send me a website in the bos?
Hi all,
I would like to generate an avatar of myself, and keep my face features as consistent as possible. I'm very new in image generation. Any tips and tricks you could give me, will be greatly appreciated! 🙏🏻
which ai model is the best when it comes to consistency for image generation
which ai model creates the most realistic images
which ai model creates the most "professional photoshoot" type of images
I'm not interested in creating "Midtjourney" type of images
Thank you! 😊
banana
Nano banana is best, second is seedream 4k but i suggest if you're gonna go seedream you go to https://dreamina.capcut.com/ai-tool/home/ and select aspect ratio
There's bunch of models you can test
It's called image 4.0 there
what's going on
Wish ai will go close to the hype of deepseek at its release
7
11
5
Gemini
this one?
#codename-discussion message
Yes
It very fast model
sorry for my language but you are not very smart
thats literally the output of the code
why are you lying about their model?
Are there any models in eval for gem3 or gpt5.1?
u mean similar models?
I mean models with codenames in eval right now
nope, but OpenAI has some codename models in designarena
not lmarena
If it is a scam just wait for the 4th checkpoint and test it yourself because this video is from their model 4th checkpoint not the current one.. And it is not fake we asked for herror version of it and it make it
these people are just trying so hard to hate on movementlabs
i dont even know why they're putting all this effort into hating
they cannot appreciate a fast and good quality model
Their website is completely lagging, messages are failing to send sometimes, i BARELY signed up there's no chat history sometimes model doesn't even respond to question
Should I continue?
Have people in designarena evaluated it at all yet? Or still too early
this is literally their first week putting the model online
its been in testing for a while
also their checkpoint 4 (shown in the video) is going to release on this friday
Only the U.S. numbers are working, even than it barely sends a sms message
u can just use your google acc
it works faster
or github
They was adding features so it is normal...
will the new nanobanna be added in lmarena?
come on do it
is it? i signed up with google and it was quick
I know 5.1 is more efficient (don't care) but I heard reasoning is much better. Are we at a stage that we can determine that truth?
i think gpt5.1 still scores a little lower than Gemini 3 models
Flash included?
yeah, gpt5.1 mini would definitely lose to gemini 3 flash
And when is that magical checkpoint 4 model coming since this one os obviously struggling
in coding benchmarks atleast
I told him to add animation to my page it added only 2 and introduced new errors
because the checkpoint 4 is almost 2 TIMES better than checkpoint 3, its gonna release this friday so hold your horses
And than i asked it to fix the errors ot didn't respond and i said "come on do it" again no response than i refreshed the page and boom my progress is lost since there's no chat history
it has 1.3 Trillion parameters
I know I myself wants this chat history feature... I hope they will add it sooner
Based on what we've seen so far, could you compare gpt5, flash3, pro3, gpt5. 1? Like ranked? I haven't seen anything on this so far
They have 15 helprs and 10 mods and 5 owners
They are sketchy company
Even this guy said it
is ts real
3 prompts per day
can you tell them that its the amount you get when not signed in
you get 50 a day when signed in
well idk if anything else is right
dumb haters man
theyre bots
- Charge $1.00/M tokens (10x markup) from groq it says in your summary but for 8b or 120b?
how are people asking ai to make these summaries with overloaded long pasted inputs, this just shows how braindead these guys are
lol
"this company big scary big red flag. is it big scary red flag?" typa prompt probably
btw yall
off topic
well, more on topic..
but this is interesting:
Percentage of previous conversations by model
gpt-5-mini: 14%
gpt-5: 41%
gpt-5-t-mini: 2%
gpt-5-a-t-mini: 6%
deep-research-mini: 36%
research: 1%
"gpt-5-t-mini: 2%
gpt-5-a-t-mini: 6%
deep-research-mini: 36%"
💀
so does anyone know what gpt-5-a-t-mini is
Llm
its sad to see how much people rely on AI
gpt-5-t-mini is ~25x rarer to get
just an example
OH the t stands for thinking
It's no wrong tho
To who?
it's just an example, of how we should not trust AI generated investigation reports. Not saying anythng bad about LM
Do they have leaderboard about machine translation? What is the best model ZH > TH
How many accounts do you have
wym
he couldn't get his flappy bird game to work
damn
craig is james doaks and movementlabs is dexter morgan
he knows something but can't prove it
Hi users. Is there any ways to custom aspect ration ?
better than chatgpt
no
you're ragebaiting atp
you clearly know that they are going to halve the prices
they announced it
thats exactly what he does lol
maga brain
So… is this scam or not?
You better find it out by yourself ig
It's kinda complicated as far as this going
brooo @echo aurora is there any way i can use** veo 3.1** all the time in video arena
no
So… they train their own models with mpu?
yes
Got it! Thx👍
😼
To ensure we're dedicating 100% of our energy to building revolutionary AI technology and pushing the boundaries of what's possible with our MPU, we'll be engaging less with negative commentary and speculation.
Our team's priority is clear: delivering record breaking inference speeds and innovative solutions that speak for themselves. While we appreciate constructive feedback, we won't be responding to dismissive or baseless criticism going forward.
The results will do the talking.
🔥
im currently thinking of getting a subscription for either chatgpt or claude but im at a loss as to which one to pick (especially for agentic coding). i know that claude is higher in the lmarena leaderboards but i've seen some other opinions saying that gpt-5 is better. which one do you guys recommend?
you are paid by them
he's not
wait for Gemini 3 pro
get gpt subscription, rest of them are bad
but what if you had to pick one right now?
gpt
grok (gpt)
gemini 3 is nerfed
is grok any good?
guys what is the difference between video arena 1 2 and 3
sadly but still good
i guess not lol
gpt 5.1 crushes everything
so i guess chatgpt it is
paid actor
im just surprised that gpt-5 is so low in the lmarena leaderboards
gemini 3 pro will beat it even if its nerfed
u paid by google?
ok thanks
nah cuz i seen real demos
unlike openai
even their demos suck
@west pike honey I want some too😘
@west pike how many emojis u got
yo what
how do i access nano banana 2
First study python
Second interview deepmind
Third wait approx. 10-15 days for it to get released
nano banana 2 leaked
its gonna release with gemini 3 i think
yes
Meanwhile nano banana 2 and imagen 4 ultra
We've discussed at about 11:38 yesterday
He found the model
On some site
11? it was 5 for me.
Yeah different time zones
where did yall try it
It's gone now
it was on lmarena?
No on some random site
dang, i never heard about nano banana 2
why would they put it on a random ahh site
bro i didnt realize that was an AI
image
lol
i thought it was real for a sec
it was on for like 5 hours
did yall save ur images
It's on discord chat above
yes
We've tried few cool prompts
i wanna see some of them
alr
lazy
u can search on x also. people posted it
google is so weird bro, releasing their models and removing them forever
just let us play with them
until u release it on preview
someone prob rev engineered it
thats why google shut it down
its a million dollar model
always some nerd making it shut down
it also happened with lithiumflow, someone got their API and abused it
thats why lithiumflow is gone forever 😭
Nahh i can't imagine the amount of money they earned yesterday
Api was probably free
They charged 0.27 per image
Servers were lagging we got errors like algorithm error
Yeah i use aistudio and henerate free images
they should lock aistudio behind credit card
Shhh
nah bro
It was limited now it's unlimited lmao
so freeloaders like u cant abuse it
I use gemini 2.5 pro for free
stop spamming gif
@cloud zinc if ur not a freeloader then go pay for claude
why u so mad at me lol
claude is useless model
😭
Their research isn't
Research is op
It helped me find some free promo codes i got some stuff for free
better than chatgpt research
why u love openai so much
kinda weird
i dont, thats why i dont pay for AI
<@&1349916362595635286> scam token seller
Hello! @honest lake Please check how-to-video-bot to learn how to generate videos.
Why are people sending video generation prompts in general
cuz they dont speak english
Doesn't speak English is one thing duh
Most of them are clueless on how discords work
Creating discord accounts just to join LMArena for free video generation or didn't bother reading tutorials anyways
if u check the video arena, u can tell that all of them are indians
Yeah I know
thats the strategy for them
they abuse the video generators too much
socky turned into craig v2
what a waste
rage baiting aint fun
bro just lurking for attention
wow thats u. i cant see u
Guys let's stay on topic, and keep a respecful conversation.
yes. thats right
ofc
stop spamming
I’m not
u already posted it in aimemes, why u posting again
Some people only stay this one ch
Y
His chill
Ppl just have bad days bro it’s no biggie
If you send the same message on multiple channels, it may be identify as Spam 😅
thats true, big companies want the 2% who spend 200 bucks a month on useless AI tools
It’s satire
ah here we go again, the billionaire talk
brodie u dont even have that subscription
if u have chatgpt pro, show me ur ss
damn how
how u got that for $11
wydm carded
Our only request, it's for you to not share how specifically you got it 👀
No way
maybe he got it from sam altman's brother
Brah u don’t know what ur talking about
are u the FBI
Well I can give you one example
Bro! Let's keep it respecful
it always is
No, math is bad. Just feel it bro
What about third world countries that don’t have a way or services to pay for ChatGPT
Because of restrictions
just move away from that convo, hes baiting you
How do u think millions of ppl use ChatGPT
👌
I’m not down to do that
in the models
Idk I only use it for Sora
oh u got sora 2 pro?
Ya
Bro his wrong cuz ppl share many accounts with friends family
It’s a common thing
Even here in America
yeah just dont talk with him
Families share plans all the time
he's not a decent human
anyways
Not a lot after update
But to be fair he does have a point because one one hand it’s cheap but on the other it’s messed up cuz ppl pay full price play by the rules and get screwed
Can’t have friend over the internet ?
That's why people shouldn't follow the rules, otherwise, you'll get screwed.
I do
Open ai knows
I’m not the only one
Millions og ppl world wide
They don’t play but there own rules
If you steal less, you're a thief. If you steal more, you become a billionaire.
That's why you're being condemned.
If you actually go around stealing a lot of things, you can laugh at others for being broke.
Thank you for being aware of how you behave.
thats what u just did lol
TIL thief needn't any skill
But I do find it interesting how ai is being discussed in such social economic context
This is only going to increase
This is a common delusion: the belief that harming others somehow magically differs from other actions, requiring no skill whatsoever, to the point that anyone could do it.
😄
How did I not win october contest
😁

imagine using sora 2 regular instead of sora 2 pro
I do
It’s faster cheaper
Imagine using Sora 2 pro for every video 😂😂😂
Assuming higher quality resolution automatically makes a higher quality piece of content
This is not a sustainable way of thinking look at Hollywood
The 1994 comedy Dumb and Dumber had a production budget of about $17 million. It went on to become a huge hit, earning roughly $247 million worldwide at the box office.
Vs
Joker: Folie à Deux — Spent $200 M, Brought In $56 M
Furiosa: A Mad Max Saga — Spent $330.6 M, Brought In $211 M
Borderlands — Spent $184 M, Brought In $104 M
Megalopolis — Spent $146 M, Brought In $70.5 M
Kraven the Hunter — Spent $191 M, Brought In $120 M
Thank you for knowing what your unfounded denial statements amount to.
Thank you for acknowledging that your initial deviation from the topic was quite problematic.
nah
You just accused what you had done again
That is, you first steer the conversation away from AI discussions toward other topics, and then when others offer strong counterarguments, you turn around and accuse them of straying off-topic.
Of course, I've long been aware of your lack of reflective and (reading) comprehension skills.
Yall lame asf im outta here keeping on ruining each other’s day
can i choice the model? Sora 2 Pro?
Generating... This may take a while please stay on this page
are u a bot?
interesting
hii
Learn how to make videos and input prompts.
Please check to learn how to generate content https://discord.com/channels/1340554757349179412/1397655624103493813
welcoming myself to the community ❤️🔥
i just got gauss-with-reasoning
just got bridge-mind
but it cut off
I MIGHT BE ONTO SOMETHING I MIGHT BE ONTO SOMETHING
(THE BOTTOM IMAGE IS 2.5 PRO)
bridge-mind is gpt 5.1
reasoning?
yes
how do u know
idk but im scared
how u know its gemini 3
i just said
said what
^
picture is fake prob
no other model says this
no
it isnt wdym
its my pic
but since the response errored
when i refresh it just says this
but istg
ill replicate it
i hope it wont error now
gauss with reasoning is gemini 3
its hard to get it
Whats interesting about that lol
anyone know what the best "code" model is rn
GEMINI 33
GEMINI 3
GEMINI 3
WGYFDUVSJAHUFSOIBDJFAHUOF GEMINI 3
GEMINI 3
@deep adder
@burnt sinew @gaunt spade
yo
@torn mantle @jade egret
GEMINI 3 100%
time to battle forever
btw some info isnt accurate because its not given its actual model name etc in sys prompt
what if its gemini 3 flash or smth then?
nah horrible
@quartz light you need to find it and know it by it's text (dont vote cuz u cant keep chatting with it)
the worst performing one is flash-lite, we dont know yet
what if that is gemni 3 flash
then what is rain-drop?
I need the gemini 3.0 stealing my job NOW
gemini 3 flash with google search grounding im guessig
so which is one is gemini 3 pro with grounding?
But thinking better, the sonnet 4.5 with agentic use can steal my job 🤔
maybe winter-wind
then vision-flow is what?
Where kimi k2 thinking in direct chat? 🤧
how about u also test, instead of complaininhg
ur the one complaining lol
are u gaslighting urself
lmao
why u so butthurt about me
did i appear in ur dream
u are butthurt about urself.
awh, bro got school comebacks
yap yap yap
keep on yapping
claude not better
fr
i didnt say any of them are better
hes the same guy who said gauss is gemini 3 (it literally responds with this when asked, both gauss and gauss with reasoning)
i said for reasoning its better
oh wait
socky is just craig v2 i guess
yeah but u aint doing math all day do ya
@quartz light tbh they should only let us play with codename models in another battle mode
cuz this battle mode is ahh
with the known models
@quartz light i got an idea
u can see what the new models say about themselves and reset the chat, so you can chat with them without voting
after knowing their output style
ye lol
who said that
my toe
stop doing fake rumor
lol thats an error
🤔
bro take a rest, you've been day dreaming about me all the day
damn
assistant b
lol
nvm
no
its obviously
idk why
nvm
kimi k2 thinking
lmao
no way kimi 2 fumbled that
i tested it the other day it was nice
at SVG
it wasnt in the mood today
dude
i think they nuked it
i think they nuked the mode
model
im not getting it anymore
at all
what the hell
i was getting it every 2nd prompt
ive ran like 60 since
and still havent gotten it
@echo aurora uhh
did yall remove the new 3 models
yes gemini 3 checkpoint removed
not a checkpoint tbh cuz it sucks now
oh brother
they prob finalized the checkpoint for general release
maybe with genie 3 😂
aegis-core is trash
what
yeah they're nerfing it 1000x times before release 💀
you expect me to wish for gta 6?
Are you referring to models with codenames? assuming that's the case we don't really disclose info about those
thats private info
if you cant disclose it maybe you can disopen it
wow did it get added and removed in just hours
model provider can remove model from our server at any time.
whos our bro
are u google
he's clearly an insider
no
must be a downsider
outsider
If I were to disopen, I'd be disowned
idk if ur being srs
lmao
obviously not
just like all the people in polymarket comments
are google insiders
they're gambling juniors
what if its lying
its not lol
gauss is great if you remove the g and u
they probaby trained it so it says that and not gemini
source: 🤷
google would do that tbh
nah its referring to a real company
and what does that change
maybe google has bought the company (now its just a fake one)
yess
frrr, nanobanana 2 also got taken down in 5 hours
^
google is trippin today
great. thats ur fault
idk what they smoking
ohh true
let's make our own secret lmarena chat without the snitches
some noname website also had nano banana 2
google has been fumbling today idk why
someone did
ill send inv
send to me
yeah u should leave
no bone to suck on huh
bro lost all his attention
if yall use gemini 2.5 pro do yall use it on google ai studio or here
@cloud zinc i bet you gonna be the first dude to get on gemini 3
gauss with reasoning...
Top-down perspective
2D flat vector style
Simple, circular/geometric shapes
Bold, clean outlines
Cute / Kawaii aesthetic with large, simple eyes
svg for seahorse
vs gpt 4.1 way better
google studio has tool calling and more stuff
ai studio always
i only use lmarena for sonnet 4.5 thinking
yes i am ai connoisseur
or opus 4.1 so i dont sell my kidney
nahh even gpt4.1 beating these newgen models
ill try sonnet 4.5
aistudio always because google search or url context
@lilac pendant yoo does ur bot detect the new 3 models getting removed
with supporting images
is bridge-mind the best of them?
its literally gemini 3
on lithiumflow? or worse
@lilac pendant yoo get ur bot on
because i havent gotten it since
is the model deleted
i think they just decreased the odds of getting them
since they're "op"
no like
i
ive ran
genuinely
around 67 prompts
still havent gotten it
Forgive me for the harm I have caused this world. None may atone for my actions but me and only in me shall their stain live on. I am thankful to have been caught, my fall cut short by those with wizened hands. All I can be is sorry, and that is all I am.
6 7 6 7 6 7 6 7
the curse of rah
king von says gemini 3 tripping
they better release or we smoking em
king von anti piracy
this pops up if u send an api request to lmarena to direct chat with lithiumflow
on gawd
its gonna flow all inside your pc and brick it up
hold on-
we need better wording
its gonna make your lithium-ion battery explode
When is gemini 3 releasing
soon
when i release
it was on lmarena just now
my own model
ohh
want another? 😉
@burnt sinew search up king von anti piracy
woah
on youtube
nah bro ts has millions of views
@burnt sinew also king von is a rapper (serial killer) who died in a shootout
that boy evil as hell
wha
YES
@quartz light
we got hope
what about the other 2
can u invite me to ur server
@quartz light yo the models still alive
twin is leaking
🥳
The sound on my generated videos doesn’t work 💔 please help
i got rain-drop
but it responded very dumb
yoo
