#general
1 messages · Page 157 of 1
thanks gehlo
Mixing match
you help me so much
u need sora 2 pro/veo 3.1 paid if ur a short creator, but that costs money
are you a content creator
if u want high resolution
u gotta pay for it
then do it on there
u kidding me
It don’t matter if they’re poor if you making funny dude to
I seen things made in PowerPoint. They are funny as hell better than some of these high-quality comedies.
What I’m saying is dude you got all the energy you got a lot of things going for you. You just got a channel it
Do something positive with it and positivity will come back
And the door will open, bro
thanks
But in the meantime, you gotta be patient to work with all you got
I expect great things from you
hope i do'nt ashame you
any help i can do for u gehlo
anything just ask may i can help
i know a lot of info about anything
may be i can help u
A true g never needs a hand, but it’s always willing to give a hand
You know how it goes
Hey, when you wake up in the morning
i'm fallinf sleep
Look in the mirror and tell yourself you the man with the plan
i will
thank you for relizeing me
youre the best and the nicest preson i ever saw thank u
gehlo
Likewise bro, good night
Until our roads meet again
we will
I’ll tell you guys what though if I was a businessman there’s a lot of golden opportunity in this market
A lot of demand and as everybody here just witnessed not a lot of supply
If somebody wants to figure out a way to be able to tap into this market at a decent price and meet these people halfway
You’ll be doing 95% better than the 95% of the AI companies that are unable to generate profit
For godsakes it’s 2025 god knows how many image updates and image generators and videos generators we got
At least give some of those old ones back to the public or at some kind of discounted rate
Ve2, Dalle 2 & 3, Sora 1, flux and the list just goes on and on. There’s absolutely no reason why some of these older models or even some of these services, can’t be at an affordable price under five bucks a month even 3.99$ a month would be something that people would be willing to pay for that normally wouldn’t pay for it
i need sora 2 invite codes please dm me
<@&1349916362595635286>
Wait, I wanna check it out before they delete it. Give me a second.
Thanks
I don’t get what they’re trading. Are they trading PC software or something?
See, I told y’all about that advertisement stuff 😝
Still need?
Sent
why do I get this in every chat now? without battlemode
Ma man
VPN..?
no
Are you usually quick when you do stuff?
it was good yesterday
I’ve noticed when I refresh a lot and start new chats. It will do that sometimes.
not that quick, but everything was great yesterday when I used it
yeah but it comes on my chat every time I send a prompt now
I am not exactly sure how it works, but I’ve also tested it when I was doing some red teaming and every time some content would come or whatever whatever it would pop up also
I’m not saying that’s what you’re doing but
It was super bad when I had a VPN
I think this is a prevent bots
i need to create a logo and i would send you an example.... ho can i do that?
hiii
/video
i just dont know butw i cant send any messages in #video-arena-2 all 3 whyyy
/generating please wait…
i need to create a simple image for logo...
WIP
Video genareto fiel
Hey folks! You should head to #1397655624103493813 for a detailed guide on how to use the bot. However, Video Arena bot looks to be down right now. Our team is looking into the issue. Thanks for your patience.
I need veo3 video create unlimited
i thought that said haiku 2.3
YO Does hailuo 2.3 have audio?
Genaretiom fiel
Reading
@echo aurora broo i dont know what happened i op[ened discord after a long time but i cant send messages in #video-arena-1 all 3 its showinhg i dont have perms likee whattt 😭😭😭
i think you need to pay for unlimoited
Check out this message - #video-arena-1 message
I've been wondering the same for awhile now
Yes
I think it's a dinosaur?
oh okaay thanks i was worrying
No problem
I’ve been generating up to 10 videos with mini max agent in bulk batches
Really loving the agent
woah bro is it freee b
Yeah, and now I could even add an audio to it
thx for the info i am going to tryu ity
You need Google subscription
And even with that you still have limits
the next upgrade to these models need to be price decrease
By the way, I’m not sponsored by them. I’m not affiliated with them. I’m just giving you guys my honest opinion.
This thing you do anything
i havnt really used veo 3.1, i think i might transition soon lol
hi friends
Almost 2 dollars a video
hello 
lol nvm
Hi
i have a question for you all
Yes got it
👋
Whats up?
Dude this kind of design with these agents as the future guys
mac vs windows?
All in one everything’s included all in one
What makes this awesome they have everything already pre-built and so easy to use
Supabase
I never vibe code in my life
Look at this cool thing that built me
Depends what I'm trying to do, so both (sry that's not too helpful
)
hey I have both too
It’s a simple token highlighter
“Wrong Name”
Music: Emotional tension (CapCut “Cinematic Chill”)
Narration:
“He said the wrong name… and revealed the right truth.”
On-screen text:
One mistake. One heartbreak.
I mostly use windows for gaming, and mac for work
what is the best open source vido gen model is it movieflow or the dtr one?
/dont show generator option
Wen
But you need a lot of compute because it’s slow
Like 30 minutes generation unless you got a lot of hardware
use sora 2, its better and free
I never realized how strange peacocks look
true but not outside of united states
are u an ai engineer
wish I was that skilled
but no I'm a Community Manager 
Sora is poor quality
use a vpn for that if ur not in the US
And watermarked like there’s no other
its very good at physics and understand your prompt
True
The Video Arena bot looks to be down right now. Our team is looking into the issue. I'll open these channels back up once fixed.
(sry everyone for all the /video attempts in the text channels. this tends to happen when we turn those video channels to read only)
OK, so the audio thing I just tried for the first
And I don’t know if I’m just doing it wrong or this is the way this is but this is kind of weird
Definitely not what I expected 😂
I used it to clone my voice last night
best AI video i seen so far
Ya really enjoying gen ai
Anmation
The video bot isn't working at the moment
So we've turned those channels to read only.
WE NEED GEMINI 3 PRO
4
5
1
yes fr
I’m always fascinated by this kind of behavior
😂😂😂😂
hey guys, new here
Sup big dog
/image to video
welcome welcome
The bot isn't working at the moment
This is hilarious
Agreed
Appreciate it
Let me know if you have any questions 
what's good my guy
/image to vedio
alright. i see the bot is down. how long has it been like this
@echo aurora hey what happened can anyone please explain me why we can't able to generate vedios ???
what model was used for this
How many videos are you generating?
Me🫤👈🏻
Yeah
Maybe daily 5
It’s down right now
About an hour I'd say, team is working on a fix asap
Because it isn't working currently
The reason I flagged you and asked you that was the way you phrased your question in the beginning
I think this problem is not fixed yet
What problem?
Correct, team is still working on it
it's saying i dont have permissions to post in the channel. is there anything else i need to do?
Nope, that's intended. Once the bot is working again I'll lift the restrictions on the channels.
If you want to make videos check out mini max agent
ok, gotcha
ola
Whats up?
Do you guys know any other platform to make ai video with free access
algum brasileiro aqui
I hope chatgpt announces some new model or something today at their event. Getting boring with only gemini 3 which will never release to look forward to

Hello, everyone new here. I am learning how to change one of my show I did into a puppet or anime version. Learning a lot and I thank you all in advance
fr
hi
hello
What if it comes out and flops?
/image to vedio
Hi
hello
It might
👋
Hope not
Why can’t I access image to video
I'd encourage you to read this blog post that goes into details about our Evaluations Product - https://news.lmarena.ai/ai-evaluations/
/image to video
Because it's currently not working
cc @rustic raft ^
Whered the embed go
"From Fire to Light"
hello
bot
Man, I’m not gonna lie I gotta admire the brazen
Just no regard for nothing but what they need to get done
Watch
I wonder what this phenomenon is called as what we’re witnessing
And I wonder how many other channels they attempted before they got to this one
Done
I am trying to find out myself, I have a 23 min video that I want to switch to a Puppet version.
i deleted them i dont want to spam channel with long messages
Hey Anyone play minecraft with me 😂😂🎀
I never played that game once in my life I don’t understand what you’re supposed to do in it
?
I just see when my sister was younger playing with it with some pickaxes or something
/vidyo
@keen beacon it's too interesting bro
seems like there would be no better timing for an announcement?
It’s too opposite events
Having a release as a big event on its own
Having an IPO event is on a magnitude of its own. I don’t think you could mix both.
Not that it can’t be done or that it’s impossible but
boost prices?
"From Fire to Light"
from hype?
It’s already saturated though, bro
People aren’t stupid. They need results.
But there could be something to it too because people are rumoring that ChatGPT six is around the corner
"People aren’t stupid. "
Sam Altman himself said that it won’t be a year for a new model
heh
i just mean announcing it not releasing
gpt 5 just came out
maybe a small incremental upgrade or something
Dude, an IPO is like the glaze capital of like BS
but larger announcement of whats to come
Hello! I’m excited to join Arena because I’m passionate about creating and exploring innovative AI videos. I look forward to connecting with the community and learning more about how everyone uses AI for creativity and content.
It took TWO YEARS
Welcome welcome! That's great to hear!
IPO glazefest
im nto saying it would release so soon after gpt 5. i just meant announcing something major whether BS or not just to generate hype
You know what you’re supposed to do when you have an IPO event
You oversell the hell out of it
yeah tahts what i mean
Like this is where you go you’re 100%
You start promising that ChatGPT will take people to the moon in two years if you have to
All good to come here then, this is a very warbly channel where they per pure accident actually manage to give some good info now and then. 😸
chatgpt 6, AGI next year, buy our shares now
We’ll talk though it wouldn’t be a bad investment
Sense it’s not gonna be able to generate revenue until like 2029 or something
So for investors, that’s taking a little bit of a risk
This whole thing is just crazy a company valued at 500 billion only is able to generate 12,000,000,009 or 10 billion expenses
Not just a leetle, but that's what those stock trading guys do anyway - as they're gambling addicts all of them.
Not me, have been buying gold in very small packages.
Well, you gotta consider everybody else around it that’s involved so open. AI is kind of cushioned.
And so far, I've done very well doing so.
You got a video you got Oracle you got soft bank you got you name it
There’s a lot of other people’s money so theoretically should keep them on their feet
Microsoft also
What did Microsoft invest in them 13 billion or something like that?
But I will tell you who is gonna lose
THE USER
He’s only got 20 bucks a month in it
I need free veo3 video create unlimited
We all know the cycle of in
Hi! I already introduced myself, but I still can't post in #video-arena-1. Could you please help me get the permissions or the role needed to generate videos? Thank you!
😂
What's good guys
That’s probably the most adorable answer I’ve ever heard today
Hello, the bot is down right now
That made me laugh because of the criteria was to introduce yourself and then you could use it
It is closed right now - as announced will be opened later. Make that a sticky Bluedog?
Ok
I see the bot in the channel, but I can’t write or send messages because I don’t have permissions yet. Could you please help me get access? Thanks!
or......
You know this behavior is also you see this in the real world
You see this with traffic you see this every day life
Since the bot is down we've turned off messaging in those channels. Once fixed, we'll open those channels back up.
I call it a funnel effect
Well sound processing done on local - byee again. 👋
ok thanks a lot
Go try free daily generation on DDG, SeaArt ...Pika still got that I think etc etc.
@somber inlet @neon wraith https://discordapp.com/channels/1340554757349179412/1400148597768720384/1432767695756988478 The video channels are currently locked because THE BOT IS DOWN. Please remember you can only generate vontento in the video channels, not in the chat channels. Thank you 🙂
Thank you very much, I understand now!
Why can't I create videos or images?
Anytime! 🙂
Sometimes I think these aren’t people, dude
But then I always remember the way people really are
See DM message from me to find your way to go and try things.
Friend please :´) The bot is down: https://discordapp.com/channels/1340554757349179412/1340554757827461211/1432782251967385681
Next time this happens, I’m gonna keep track to see who the champ is
hello
Who keeps going after the most deletions?
Hello
My money is still on the soccer ball
Pessoal não está fazendo mais vídeo não?
done
Has anybody ever watched this documentary?
Very interesting
They say world history, a lot of events that occurred in world history 2/3 of it could be devoted to stupidity
1?2?3?😭😭😭
Do you guys know anybody or even a single intelligent person that could define what stupidity is?
Anyone did some true tests //coding// with GLM 4.6? How you compare it to Deepseek ?
It’s been getting a lot of praise, especially from people here that use
what people? people also loved deepseek on week 1 then PUFF
Just the chatter I’ve been hearing on here from people that use it. They really seem to like it.
Well ya that’s why the number one benchmark doesn’t really mean anything to me
Because it’s always gonna be the newest model
I mean, what kind of result is that if the number one model is always the newest model
hello
Hi
I can't generate the video. What's wrong?
I don't know on what GLM is sitting on but if you ask him what is the latest laravel version he know he will reply 11 (that was true for 2024) now we are on version 12.. And laravel you know is not unknown php framework is THE framework
To be honest, dude I been around the community for a little bit in quite a while. You could say I guess.
And one thing I picked up on is that everybody uses AI differently and judges it according to their own merit
For you, it seems like that coding is a very important quality that you appreciate
Other people use it more as a companion or just something or somebody to talk to and could care less about the coding
It’s really hard to say because people just got different interest and how they use it
The bot is down roght now and the video channels are locked at the moment. We´re working on it. Thank you 🙂
I think like the real hard-core enthusiast really appreciate the benchmarks (but they lose touch with what regular every day people use it for)
Hello
Wtf
It’s down bruh
How to generate video guys?
😆
Usually this si the process: https://discordapp.com/channels/1340554757349179412/1397655624103493813 but the bot is down at the moment :´)
Just saying that both deepseek and GLM are outdated in terms of frameworks, and if they are outdated on top frameworks it means that are absolutelly not what they are advertised as "#1"
And sonnet is sitting at data from Jan 2025
I think you’re judging it too harshly
But yes.. you said it.. if you need virtual girlfriend you are on the spot 😄
Use the uploaded image as a reference. Create a realistic romantic video where the man from the image is singing the Hindi song:
“Hum Tumhare Hain, Tumhare Sanam Jaaneman Mohabbat Ki Har Kasam Ki Kasam Na Judaa Honge Hum, Hum Tumhare Hain, Tumhare Sanam.”
The man’s lip movements should perfectly match the lyrics of the song (no actual song audio, only natural background sounds). His expressions, eyes, head, and hand gestures should move emotionally in sync with the meaning of the lyrics — showing love, affection, and romance.
Along with him, show the woman dancing gracefully with him. Both are dancing together to the rhythm of the song. The man moves around the woman, circling her gently while dancing. Their facial expressions should be romantic, and their body language synchronized with the emotional meaning of the song.
Use natural ambient sound effects (like wind, soft footsteps, slight fabric movement, etc.) instead of music.
The lighting should be soft and warm, creating a romantic atmosphere. The background can be a beautiful garden or open scenic place, softly lit like an evening setting.
Training models takes time, so how are you gonna train a model in February 2025
And then plan to release it six months later and it’ll still be February 2025
You can’t go back in time lol
Plus, if it’s true, they say that they’ve been trained on all the Internet what more are you really gonna add to it that it already doesn’t have just minor little things here and there
I'm because.. you know .. AI will destroy the devs .. will take our job... maybe but wont be alive to see it 🙂
Of course you are right and I agree with you but.. ok for all opensource models to be lagging but "premium" one like sonnet / opus ..
Lmarena???
@idle turtle A friendly reminder that english is the only allowed language in the chat channels 🙂
I think what we’re seeing with these layoffs and all this talk about AI taking peoples job. I think that is being used as an excuse to clean house
hello
haha absolutely 😄
And they’re only reason I say that because of anybody has used AI in the last year
The only people that are really gonna lose their job is gonna be like things that could be replaced by like mechanical things
Ok bro, haw are you?
Anything that could be replaced with a robot that’s probably a more concern
Doing good, thanks. The bot is back!!
#ai #techjobs #jobmarket
From Tesla and Microsoft to Google, Amazon, and McDonald’s — replacing humans with AI is going horribly wrong. Companies rushed to automate jobs, only to face costly failures, service breakdowns, and public backlash.
Over 90% of AI projects deliver no measurable returns, while firms like Klarna, Duolingo, and Spo...
Thank you all for your patience! The Video Arena bot is working again.
Yes
YES
YES
@echo aurora thank you all
Dude, I was in California and I went to the little convenience store and they had a robotic Jamba Juice machine
And it was just an arm
We are cooked

Here’s the exact same machine
You should see the waymos driving around. It's pretty wild at first.
It tripped me out
30 years from now on robots are gonna take over

It’s just weird seeing it in like the real world
Just kinda a eerie feeling seeing that arm
Soon it's gonna be waiters
I don’t believe in any of that, dude
I think we underestimate ourselves as humans
We’re not some kind of peaceful species
We will destroy the robots before we let it screw us over. I guarantee you.
All it’s gonna take is one mistake and it will happen when a robot accidentally or for whatever reason ends up hurting or killing a person and God forbid. It’s a child.
Oh man, I already can picture the outcry and outrage
You would have to be really naïve to think that there won’t be any accidents with robotics and people interacting
It’s not a matter of if it’s a matter of when
There is a balance in the universe and there’s balance in a lot of things and what is the balance for intelligence?
Stupidly. And believe it or not this may sound contrary, but stupidity is equally as important as intelligence.
Here’s a simple test to prove my theory: if anyone here thinks they’re intelligent or believes they are then define what stupidity is, and explain it. Surely, an intelligent person should be able to do that.
Anyways guys, I gotta go for a few days. I got work. There’s a pleasure everybody. Is you guys later
Have a good one!
anything interesting from the openai livestream?
Once I generate a video, where can I see it?
It'll appear in the same channel you generated in. You'll be sent a Direct Message + you'll be pinged in the channel as well.
Hi, am new to discord. so in the video-arena-2, I have given a pic and asked it generate a video. In that group I could see other videos, but not mine. I did not get any msg
Hi, any one can help me with this
Taking a look.
It doesn't look like you've prompted the bot. You'll want to use either /video or /image-to-video in order to start the bot. More information can be found in #1397655624103493813
hello ! I'm new here
I have a ecommerce shop
Does anyone know a place i can make a 23 min video from the 5 sec videos I got from here. Also How many 5 sec videos can I make here or is it unlimited.
It's going to be 5 generation requests per 24/hr
Cool
Thank you for your reply that help
No problem
Keep me updated if you have any other questions.
what would say is a better gen even tho they are different scenes?
Cool video tho
remove this model, it's now gone from the api
hallo
rip 3.5 my beloved
It was the first actually good coding model imo
best agentic coding ai model
7
14
3
sonnet 4.5
Looking into.
In my opinion, no, since it was a very good model of natural writing and coding.
Appears to be working fine.
Must just be the fact that it's somehow hosted on a special server.
may have been rerouted to another model
keep in mind sonnet 4.5 also claims to be 3.5 sonnet most of the time
oh wait official anthropic docs say it should return an error, maybe lmarena is using another api that still has it up?
even then, it will likely be pulled soon
Hi
yeah vertex doesn't mention it anymore but bedrock does
probably will be pulled soon
Any Gemini 3 news? November is fast approaching
Nothing really matters in the LLM space except for genuine Sota frontier progress, we can all settle into the grooves of our models once it truly plateaus but until than we should demand real improvement, the last genuine push forwards in the LLM space was 4.0 opus and that was a while ago
have a nice day everyone
GPT 5 was a titanic nothing burger
My custom instructions are too much of a sketch to share them tbh. But currently, talking about gpt5-thinking, you basically need to instruct it to use web search and consider plenty of sources. Also to use all the other tools as much as possible. That directly increases test-time compute, gives you deeper search and doesn't clash with their enforced updated instructions telling it to be concise. Also indirectly instructing it to be verbose - like saying to to not leave out any details. But don't directly tell it 'be extremely detailed' or similar - this will clash with their system work and won't work it will ignore that. I'm also still instructing it to output tl;dr first at the top, this works great so I can see what it came up with immediately and then scroll down if I need more info. And manually choosing 5-Instant for prompts that are immediately clear to be basic/easy.
for prompting images, do i have to make a new chat every time its a new image or is there a way to remove old images from the context?
You can just continue your chat further down
with new prompt
@ocean vortex vote #video-arena-1 message
someone vote
i need to know the model
veo-3.1-fast-audio vs. veo-3-fast
thanks
i see really a huge difference the left one is so good
most models see what the previous image looks like which obviously is massive bias
yeah left one better
i think, maybe theyre reading the text?
Yeah that's a quirk of lmarena. This is even a bigger problem with LLMs when flooded context biases it towards style or even identity of another model. I've highlighted this repeatedly to @echo aurora but it seems lmarena team just wants this to be like that and are prepared to deal with side-effects... Interesting choice but oh well...
Good news is that image generation does not seem to break with it, but you can reference earlier messages so they do have it in context as well
Funny.. GLM (Cline+VS ) found and fix laravel/js/api issue that 25$ on Opus and 4$ on GPT 5 failed to do.. I was sceptical for GLM because.. the hype around it and the fact that reddit is under bots spam about this model . So yes.. @keen beacon
okay i can just make a new one each time, its just sometimes i have longer texts that i might want to reference that get flooded away. oh well
I'll be able to share more when I can, but we are working on an update to this behind the scenes. cc @keen beacon
hello everyone just here to learn more about ai 👋
hello 
will be a lot more convenient, looking forward to that
I haven't confirmed what that change will be, just that we're working on something related currently.
Lol we added a new model on Friday
Chatgpt 5 💀
I mean not video gen only

related to making things more convenient with images im sure that whatever it is will be an improvement
Would like to see new coder models or reasoning ones
Offtopic: Anyone facing VS Cline Long Pending task stuck ?
Do you have an openai that I can use for roleplaying
For janitor ai
The default LLM is kinda lame

@polar niche
where is orionmist
what channel is this where to get this notifs?
not my pic originally
It's from a bot in a different server.
if anyone has the link to this please send dm i would like to be always uptodate
Me too
The team is aware of some issues with all models erroring out. Team is looking into and working on a fix ASAP.
people ddosing trying to spam to get lithiumflow?
lol
i saw in twitter nano banana 2 is coming with gemini 3 is that true 👀
idk (maybe next year?)
but do you like boardgames?
i just recreated one: #ai-creations message
-# with magnet mechanics! (https://boardgame.tips/shogun-ravensburger)
Unlikely, at least at the beginning. They still seem to be pushing Imagen models for their low cost.
Should Claude 3.5 Sonnet be removed from LM Arena?
5
9
4
No, it was so, so good at many things!!1!!1
Anyone have any examples from lithium? I’m super curious what kind of results it pushes
I've got one.
It was from a while ago when Lithiumflow had just gotten added to LM Arena (I believe).
Would love to see!
Here you go: #ai-creations message
It's an accurate recreation of Tetris.
Super cool! I’ll check it out as soon as I get home
what is this model, and its creativity looks good
hi !
can i get the link from the server? i would really love to be up to date!
??????
why is it mysterious
@ananas grrr
erine exp?
I don't think I have invite perms to this server sorry to say.
ah, fair enough
if possible, can you atleast giv me updates?
I am gonna find you Raptor.
u a part of it?
yeah
yep, don't know when it occurs, and the writing is gemini style recognizable.
who runs it?
names a person thorne?
Sounds good.
something like lena, thorne, aris, dataslate and slow resonant hum.
Heh.
Just go ham and keep spamming the same thing over and over.
Or, if you wanna go in depth, you could try taking a look at the “Global Events” thing in “Network”.
On the Inspect Elements window.
Alrighty.
.
hm
LMFAOOOOOOO
No one has ever said "AI" in an AI server. Ain't no way. 🥀
no its the fact that there are so many people that said ai
discord was like
nope
heck you
then it just gave me nan
Makes it look more like Discord was just lying and trying to cover it up.
wym finding?
I think he means having a higher chance of getting anonymous models.
finding bots like phantom-1017-1 on battle mode
Alright.
how come
half life 3 before gemini 3
is lithiumflow limited to just webdev arena or is it on normal lmarena as well?
u gonna say? cuz in my experience it nailed my benchmarks
( the benchmarks are custom made based on the tasks i do often )
rn? idk
anyone knows which model this is?
Not quite sure, since it seems pretty stubborn to reveal its exact model name or the company it works for.
My guess, though, is that it's likely a very low-quality model.
or this one
That one is an Amazon model, specifically Nova, I think.
im trying to find out if there is lithiumflow on classic lmarena
but probably not it seems
It also seems like it isn't on WebDev since I haven't gotten it yet on there.
I liked the previous raptor llm from last week, must be a new base they are using
only got it once though
first time seeing this one too
I remember getting that model once. I don't think it was actually too bad.
i typed test. the other models responded with simple sentences and this one straight up just developed an app
lmao
all models i tried easily tell who their creator is..
Maybe it thought you wanted to test a website, heh heh.
when does https://lmarena.ai/leaderboard/text get updated
View rankings across various LLMs on their versatility, linguistic precision, and cultural context across text.
oh yeah thats a long delay, I didn't notice that
That might mean something
@hardy lion
Oh nice.
It must have just gotten added back because I wasn't able to get it for the longest time.
Oh wait, apparently this is a Microsoft model. I just got it and it told me that it's a Microsoft model.
eww
did not get it yet
been trying to get it for at least an hour
Me neither.
I need a code for sora 2
is it actually back or did we get trolled?
I think we might've gotten trolled, but who knows?
I can get you one.
also from reddit?
Just DM me.
are these notifs from this or another discord?
cuz i cant find em anywhere and always see everyone posting them
I'm unsure, but I'm very tempted to see where this notification comes from.
It could prove very useful and beneficial for me.
i lithium really back
No one knows for sure just yet, but apparently someone reported that it is back. However, I'm not too sure I believe them.
maybe we been trolled
probably not because i tried like 40 chats already and didnt get it
😭
i think we got trolled
wheres the screenshot from tho
someone on reddit sent it
where does one even get access to the bot?
i keep seeing screenshots of it everywhere but couldnt find it
He's honestly just like me.
oh thats not
wtf
hai everyone
Olá
Oi.
@hardy lion yuh clayton
😢
well
Apparently, Gemini 3.0 drops tomorrow according to some Reddit post.
couldnt sleep as you can tell

btw i talked to someone
the thing we were trying got patched already
lol
Go figure.
I'm not much of a sleeper either.
Yeah, that checks out. Otherwise, it would have worked when we were trying it.
All they had to do was not publicly post about it, and then it wouldn't have gotten removed.
Yes custom bot
it drops tomorrow?
Apparently.
according to u?
I can't make any promises though, since I'm not a Google developer.
what are u basing on that its dropping tomorrow?
No, if you use your eyes, you can tell that I'm talking about a Reddit post, not me specifically.
Finally dropping?
Correct.
Yes reddit knows what they are talking about 🤣
According to Reddit (the most unreliable place for information on Earth).
the post has links fake ai twitter hype guys. and has 0 upvote. why u trust it
They say this everyday
To be fair, there's a reason why I said "apparently" and not "Wow guys, Gemini 3.0 is dropping tomorrow, I can't believe it!!!1!111!1111!!!"
why would u bring it up from a random post
Well, just because.
Conspiracy is what it is
Seems like it.
This post is joking lol
ikr, he still believes it 🤡
For the record, I never publicly stated that I believe it, so you can't put words in my mouth like that.
I don't think it will but it'd be kinda hilarious if it did
True.
its 100% releasing mid november as experimental.
Hm.
they will do general access full release in december, so why bother early.
I suppose I'll get a nice birthday gift then.
We know it'll release this year for sure
Yeah.
but it wont release mid-late december. google always release in first 2 week of dec
lithium come backkkk we need uuu
D:
dont know how thius works here?!?!?
You don't know how what works here?
to gererate a animated video
Ah, I see. To generate an animated video, you're going to want to go into #video-arena-1.
in measn people write their request and than the ai makes them .... my request look like im talkin 2 ma self
To create a video, you'll want to go to the channel I just linked in my message.
ok but when i write my request . nothing happens
That's because you're not supposed to just write your request. You're supposed to use the video command, which is /video, and then your prompt.
ahhhhh gotchu . ok but ive a image that i want to use for ma request. whats than ?!
If you want to use an image for your request, simply type /image-to-video followed by the image you want, and it will animate the image you send.
people should dance and palm trees weeping smooth in the wind while the sea is sparkling, meanwhile the clouds and some several seagulls are passing by /image-to-video it this correct !?
Correct. You'll first want to insert /image-to-video, then your prompt, and make sure that you have your image uploaded and attached to your message.
the person infront has three hands
Two good hands and one really messed up arm and hand.
the persdon iks a mutanz
At least the guy in the front can have a nice night at home while working on stuff with his other two hands.
thew person is a mutant
mutant
ok i try it now thank very much !!!!
It's my pleasure. I'm glad I could help you.
didion't worked ....desperated !
Did the bot give you any message, or did it just straight up not work?
no message no reaction . is like im only chattin there 😩
im in video arena 1 time 5:03
Is that how much time you spent waiting, or is that the amount of time the video was made in?
5 minute long time
You'll also want to make sure that your message is formatted properly. That way, it doesn't end up getting ignored.
no this only if you look up for it . forget to writze that i on ly need 10 seconds#
Ah.
ive no experience here
Which is the exact reason as to why I'm trying to help you.
It's all good. You don't have to beat yourself up over it. Everyone makes mistakes. That's what helps us grow.
It's really my pleasure. I'm glad I can help you feel a little less lost.
first type /image-to-video and press enter key
do u see this
yes
then add the image
ok i did ewverthing ...but after pressing enter ...whoooo where is it ?
u need to do it in https://discord.com/channels/1340554757349179412/1397655695150682194
yes id ...how long it will take in general to make it ?
It'll usually take quite a while. However, the bot will notify you in your DMs once it's finished generating your videos.
aaaahhhh ok .but i cant see my message that why i wuz so confused . ya know
Yeah, that's very understandable. The reason it does that is to notify you that it's going to create your videos. Messages aren't allowed in that channel, so it deletes it.
alos my request.
right;?! i mean till he is done withe animatiion, than it will pop up again
yes

Yes, exactly. From that point, all you have to do is simply wait. Once the bot is done making your videos, it'll notify you in your DMs.
From that point, you can view the videos that the bot made and vote for which one you think looks better to you.
thank you very much darkness for all the help and the info ...also thank u socky3635....shoutouts from Berlin germany🤙 ✌️
It is truly our pleasure. We hope you enjoy your videos and that you have a wonderful time in this server. Shoutouts to Berlin, Germany.
I'm from Washington.
ma man
Please give me sora2 code 🥺
I gotcha.
@granite heron
mmmh nothing happened.... i try it again
Ah, that's a shame.
Well, I hope they get generated this time around. If they don't, then rinse and repeat.
maaaaan they made it ....not as imagined but its ok 4 da 1st time
Well, all that matters is that they both got generated. I'm glad you're at least happy with the results. They may not be the best results, sure, but they're still decent results, I imagine.
hi guys
Greetings, welcome to the LM Arena server.
thanks
My pleasure.
I need sora2 code
i have solution
Yes tell me
what's the max limit per day
how to off slowmode in video arena
Why is there no video in my generated output videos?
hi
hello
Hello!
Hellooo I have a question I'm stuck at generating mode (using Claude) what to do ;-; already cleared cache and switched devices abzbxbnx
hellow
hi!!
what to do whshxhdf
saya dari Indonesia
Bot stop
restart it
no bot for now
bruh bigboy what are you doing i dont know any thing about what you saying its just placebo i think
okay bruh
vote on my poll i want to see why grok 4 is lower then gpt 5
💯
I'm telling you before anything that I only use gpt 5 pro, for quick prompts i use gpt 5 extended thinking
hi all
hello
hello
HI there all.
I hoep you all are well.
Could anyone or our host can tell us can we use the pictures and videos for commercial use ?
It failed at providing as reliable and usable product as chatgpt and it also failed massively on marketing ofc. Ever since Musk associated himself with Republicans and publicly went against democracy in Europe (by openly promoting and aligning with AfD)
gemini 3 releasing today, im putting good energy fo it lol
Grok4 model by itself is actually plenty performant enough to be one of the most popular ones, but it’s all those other factors. Model is not everything.
Everything else staying the same, their model would have to be better than chatgpt by A LOT, to challenge it in popularity, in my opinion.
is lithiumflow back?
it makes sense because of 30 October(3.0)well lets just hope
Ah, that was for yesterday, but alright, heh heh.
Also, no.
but in my timezone it was today you sended that msg
only for 5 second then get removed
Yeah, same for me.
udwsjdsjd
That's why I said “yesterday”.
Yeah.
Anyway… jendjfjendkwoeojrnskxndndfn!!!1!1!11
!!!!
I think that's either a pattern or just meaningless things/accidentally pressed release lithiumflow😹
hopefully a pattern
Could be.
Maybe it means that it'll be released soon.
But, it's best not to take my word on that.
hopefully
I heard somewhere in December for a full model release, and a November early release.
So, I'll have a splendid surprise for my birthday. 🤑
btw Logan Killpatrick said this hinself,it means that they started talking about Gemini 3 publicly
In my opinion, I think the right one actually looks better.
It looks more modern and almost like a glass design.
Well, that's good.
That likely means we will be getting Gemini 3.0 either in November or even December.
we're not getting it tomorrow look at the betting markets
2% chance. an insider could 50x their money
betting is not good
and betting not always accurate,nor the rumors and my word
well if you believe its coming tomorrow you should 50x your money
btw guys do you think gpt 5 pro only using base gpt 5 with more thinking time/juice and with Tree of Thoughts/Parallel Thinking....and debating himself
so...should we mimic that into any model we want by using tool?
imagine Minimax M2 Pro
genuinely true. there is a huge incentive to be close to the right price
or GLM 4.6 Pro
its based on the leaderboard and it hasnt been updated since october 16th
@lapis sage @rustic minnow @crisp plume @mortal venture Please head to #1397655624103493813 for a detailed guide on how to use the bot
hi
Test of today.
Kling left with mudface and not good prompt adherance. Top right Veo which came out of the same test. Bottom right generated on WAN website for comparision.
Winner: WAN.
See above, my winner today were WAN, with VEO not bad. Previous winners using other start images have been DDG [Deep dream generator] and SeaArt.
And not reported, Sora 2 have won once also with a scene of a woman walking trough a park at sunset.
hello
Have you guys used O3? It seems better to me than GPT-5-High ...
u not in us?
how much does it say for today gemini 3 release
gonna put some money in it
U new?
crazy that gemini 2.5 pro is the best ai and 6 months old, and gemini 3 is just going to make the lead insane
at least on the leaderboard lol
there's no market for today. 31st is earliest option
no im not in US
strawberry man said maybe gemini 3 today
No, just not that active. 😅 Is this common?
o3 predates gpt 5 and is better under specific circamstances
I thinl everyone here has used o3 btw
My goodness! There's ppl from US of A here. [Hides].
thats why i said on leaderboard >.<
although claude seems to of overtaken on some things
Ahhhh us citizens run Aibo, run!
is there any platform is good as ai studio, with context window tracking, temp and p value etc?
i noticed it can handle over 300k tokens without crashing a browser now
The leaderboard is garbage
local interfaces, maybe?
the leaderboard is really a judge of how useable a llm is and not iq level, it serves a purpose ;_;
like people are probably voting based on how paragraphs are structured more than fact checking responses unfortunately D:
Voting based on vibes. Thats a thing already
hell, i unsubbed from chat gpt, the formatting was just worse than gemini and gemini is free
but i mainly use llms for mars simulations so i gotta read blocks of text and im not coding or anything that actually requires a good ai -_- (at least not much)
HI
Is lithiumflow back on lmarena voting?
What?
Ok
I hate discord bruh
Yeah i give up
Anyone ?
which means its 100% about to hit ai studio #FactCheckedByRealAmericans
chatgpt 5 way better than gemini
yes and no
Bro gemini is way better especially in COST it's fully free to use in their AI studio and it's cool more than gpt 5 which is very limited everyday
gemini still has better natural speech, yeah gpt 5 is better at coding, math, literally everything else
die again
but it costs bro
if we are talking in something free gemini 2.5 pro wins in all conditions and I hope gemini 3.0 would be free as well cuz it would be a bomb
The situation in the future will likely change significantly, as Google has officially started to tighten the permissions and quotas for the free tier
technically there is a limit in ai studio, ive only reached it like 3 times, but if you do back and forth convos for like 8 hours straight you can hit it lol
LOL but at least their limitation is very good compare to gpt right?
yeah gpt limits are annoying as hell, another reason i unsubbed
Oh
Exactly !
Very helpful
Fun fact: If AI Studio has an ongoing A/B test, its quota is separate from the normal quota, and it is usually larger.
oh that might be why i hit the ai studio limit recently, yeah we might be cooked
Wonder if you let this run a few days your returns lol
cost to impact ratio , yes gemini wins ... but overall gpt5 > gemini
even tho their gemini cli is cool sm in peronal uses like in a your code for a specific tasks or Agent
wheres this a b testing at? i dont see it
@echo aurora @exotic slate @placid flame
Thanks for you report!
if you are actually working for a living and your coding something, you will be willing to pay for the time gpt 5 can save over gemini
Me personally I spend the extra 30 seconds to not have to use paid apis
I don't think this comparison will still like this with gemini 3.0 for a while, okay as we saw in some benchmarks that gpt 5 high wins on gemini 3.0 in 5% like as I remember:
gemini 3.0 50% 10/20
gpt 5 high 55% 11/20
and if gemini 3.0 still as free as gemini 2.5 it would be a strike !
the difference is way more than 30 seconds -_-
It only triggers within AI Studio, with models that have the 'thinking' capability enabled. The probability is extremely small, but you can write an automated program to retry and capture it
True
Also true but man I am broke like I still a school student and yeah I really apperciate free things 🤣
Hey guys! I have been finding allot of inconcistency in my image advetisement creations. What would u suggest for me to do so that I can assure maximum accuracy? and what tools would u suggest?
How so
I copy from 4.5 sonnet thinking and paste the instructions it gives me into gemini cli
I heard that lithumflow has been back is in LMarena is that true?
Same i asked a few minutes ago here too
But idk
I do these stuff sometimes 😂 ah man poors lifes
why i got failed generation
die again
u know?
me?
if you can one shot some hard problem with a 1% better ai its worth $20 to save you a day of retrying with a worse ai that isnt capable of that level of problem solving
nope,i talking to Vova.
LithiumFlow returned for a few hours, but it seems no one really managed to capture it again, and then it suddenly went offline once more.
wait I have a question
is it true that when gemini 3.0 released it wouldn't be as good as it is in a/b or LMarena version?
no
because I heard it would cost a lot so they will have to downgrade it?
so it is a lie?
same conspiracy with every release
It can't be called a lie, only 'incomplete.' You know, many people can't distinguish which 'checkpoint' is the 'Flash,' 'Pro,' or some other unknown-positioning model.
So it is just a lies and propaganda ?

