general

[echo aurora]

It's okay, we want to hear what you all think.

Started a thread here btw - #1438733849235558480 message

[sullen quest]

????????????????????????????????

[echo aurora]

not trying to make you all not talk about this in this channel or anything, I imagine others will want to share over the next couple of days.

Can you share a bit more as to why? Same thing @hasty thorn

[sullen quest]

I see, without the retry button, people who don't know how to go back to the main menu will not be able to continue using the ai

[astral blaze]

Is removing functions and options considered unlimited now? Lol

[obsidian cargo]

It's certainly inconvenient

[sullen quest]

how does this block you from actually using the ai? you can still create new chats

If I have 2 ways to do something unlimited, and someone removes one, its still unlimited

[obsidian cargo]

Yes but it's Annoying and leads to cluttered chats you have to delete if you don't want a super long chat list

[astral blaze]

So having limits is now... Unlimited?

[echo aurora]

Esp without something like a folder system

[astral blaze]

You know what let's just drop this

[sullen quest]

ORZ are you trolling

there's still no limit to use of ai on lmarena

[astral blaze]

Oh by the way since you are here the cloudflare thing is not letting me scroll down on my chat history

you load maybe twice and you hit a cloudflare capcha. Two more and you are rate limited

[hasty thorn]

I don't like that update, please put the Retry button back in.

[echo aurora]

Any chance you'd be able to share a video of this happening?

[astral blaze]

I'll try to when I have time but basically it happens when you keep scrolling down the list

I understand this is to stop model abuse but any chance this excludes looking at your chat history

but that might just be a me problem cause I have a looong list of history

[lyric yacht]

is there anyone looking for a dev?

[proven grove]

Same here. When I open the list and work on it, the Error 1015 pops up and I'm temporarily disconnected. So even though I've removed a lot of the list, I'm still temporarily blocked from accessing the site, so I keep the tab closed.
Is this phenomenon related to the removal of the retry button?

[astral blaze]

No its not a retry button thing but it's been this way for a long time now, I can't look at super old chat logs because of this.
But all of the sudden it seems fine now

[proven grove]

It still happens to me

[astral blaze]

yeah it's bad

I think it's especially bad when it's peak hours

[proven grove]

It happens when I turn the page about 5 times

[echo aurora]

models got error and returns nothing very often
I could use more clarification on this, as the retry button should still be there if the model errors out. What does the "returns nothing" look like? There just isn't a response, but you're able to vote as if there was?

[dull jay]

yes

[echo aurora]

We're aware of an issue where the model will give out a part of the response, then voting options appear. However, this isn't very common. And I don't believe it's been reported that 0 of the response will "generate" before voting appears.

[grim granite]

Same here... now i come with Error 1015.. this is a bug or something?

[echo aurora]

You're getting this commonly?

[dull jay]

this happens on models from OpenAI mostly

[echo aurora]

Sorry just seeing this. If you open your history list, click on a new chat you're getting this error page?

[echo aurora]

You're getting this 1015 error too?

[grim granite]

yeha, as the retry button is gone, i had repeat all the process and the page show me the error

[proven grove]

If I open a new page, open a previous record, or even type a prompt, all of the activities open the tab and proceed about five times, error 1015 occurs

[astral blaze]

Actually I made a thread in feedback

You guys can post there if you want

[lofty frigate]

Ummmm?? Did they remove the retry?

[echo aurora]

In the Battle mode, yes. We're collecting feedback here #1438733849235558480 so please let us know what you think.

[lofty frigate]

Ok

[tiny violet]

Genera una imágen de un toro ciendo entrevistado por un hombre en un podcast

[keen beacon]

wait

there's chat gpt 5

wtf

[elder solar]

5.1

[keen beacon]

oh

damn

[elder solar]

it looks like it can make good vsg designs

[keen beacon]

@elder solar

do you have error message retry after

4th try in direct chat

because i just did

[elder solar]

use poe

or yupp ai

[keen beacon]

im using 4o

im testing rn

[jade egret]

mobile

idk if it still here

might be gone

[sharp crane]

Hello

[chrome tartan]

the retry button dissapeared

[whole sundial]

<@&1349916362595635286>

[river hamlet]

Changing modalities will start a new chat session. Your current session will be saved in your chat history.

Why is it no longer possible to send more than one photo in a conversation with AI?

[echo aurora]

We made some changes to the Retry button recently. If there is feedback you'd like to share you're encouraged to use #1438733849235558480

[echo aurora]

Oh this does look like a bug, I just tested and when you close out of the window it should keep you in vision.

I'll go ahead and raise to the team, I don't believe this is intended behavoir.

[chrome tartan]

Here's my feedback

Terrible, please give it back

[astral blaze]

By changing they actually mean just yeeting it out the window

[stuck orchid]

gpt-5.1 = gpt-5.1-high?

Or gpt-5.1 = gpt-5.1-chat?

I think gpt-5.1 thinks less than gpt-5-high

[verbal nimbus]

GPT-5.1 Instant the chat version

[verbal nimbus]

It's supposed to be better at allocating reasoning effort

Oh if you mean for LMArena, then idk

[steep sandal]

With such a restricted generation limit and now the inability to retry images, there aren't many reasons left to continue using LMArena

[dry nacelle]

There is a problem with the website: when sending two or more messages, a notification of an error appears, the conversation is deleted, and you have to create another one, but it malfunctions again after two messages.

[solar kettle]

arena such a good application all i need is here so satisfied aboiut this web, i hope this website will get more successnes in future

[river hamlet]

yep(

[river hamlet]

Thanks!

[dry nacelle]

Is there a solution to the problem I raised, please?

[river hamlet]

There have been a lot of bugs in the last 3 days. Before that, everything was working perfectly and was stable.

[dry nacelle]

Yes, I see that, but the problem started for me about 8 hours ago.

[river hamlet]

Yes, with the photos, sometime yesterday.

[limpid rain]

are there any models provided for free by an api with decent image/text detectoin

im looking for something that i could sub in for tesseract OCR (time delay is not an issue, as long as im not being rate limited too hard)

[polar fiber]

Is there anyway to upscale images?

[steep sandal]

Not that I'm aware of. Given that, your best option is to use models like SD4, which natively support 4K

[dry nacelle]

. Is there a solution to this? The problem is constant at the start of every conversation.

[robust sluice]

yeah.. this happens a lot when I finished the prompt and enter, then everything is gone

so I have to copy the prompt everytime before I send it

[sudden shale]

why generated videos has no any sound?

[drifting crow]

https://arxiv.org/abs/2510.13928

[modest prism]

Is "gpt-5.1" the thinking variant or chat (instant) one?

[knotty fable]

Sound is a random option, sometimes one AI with sound generation get chosen for a generation, but not always.
Personally I rather prefer without. There's been only one case when the AI understood the scene for me, and added a narrative voice which was quite clever, the voice were in american english so it had to be removed anyway. [Some clips from here is used in my script film version of my films as placeholders before I do the actual generation with frame / sequence conditioning, or even with animation.]

[sharp mirage]

i think the instant one

hi

i am trying new gpt5.1 😄

to now gpt 5.1 is good at powershell codes

[knotty fable]

Hmm, AI start to understand my music - ok, have to admit some progress in that area then.....

Also sound quality is now acceptable with one of those at least.

[sharp mirage]

what ai do you use ?

okay nvm gpt 5.1 is bad at powershell

[knotty fable]

Was that for me Ebil? Unusual one, I fed a framework track of mine into https://musicextend.com/ and whoa! The result was actually very smart.

Besides Aiva, which is built on music theory, I have not seen one AI figure out the score and built on it adding details I had not thought of.

[queen veldt]

Finally i can expand the chicken ringtone

https://youtu.be/HfWPs14ehPk

So it can ring longer in the morning

😂

[knotty fable]

Or join the trend of doing a one hour version! 😸

[pseudo fable]

@echo aurora 안녕하세요, 왜 재시도 기능 키를 더 이상 사용할 수 없나요? 이 기능 키를 비활성화하는 이유는 무엇인가요? 작업 진행에 심각한 영향을 미치고 있어요.
Hello, why can't I use the retry function key anymore? What's the point of disabling this function key? It's seriously affecting my work progress.

[hallow igloo]

https://tenor.com/view/breaking-bad-having-a-job-having-no-job-jobless-job-gif-14714024496373948723

@keen beacon

[autumn glen]

testing ai

[pseudo fable]

The retry function key is disabled, so I can't even rerun the program.

What exactly is the site administrator doing?🙀

[fleet lintel]

Why so aggressive? Tone it down and politely request. It's not that hard.

[pseudo fable]

I'm sorry, this unauthorized alteration has made me a little agitated. I meant no harm, I just want an explanation.

[half mist]

Btw

[fossil fable]

waiting for 5.1 to eat the logic of lmarena live as it proves a few things

[fossil fable]

What

What did bro even say

[fossil fable]

✅

(praying any ios app would run on 15 as that beats every other official app exc deepseek)

[balmy mist]

Can’t wait for gemini3 api

[zealous sparrow]

Today the gemini 2.0 image gen models are being deprecated

[halcyon nimbus]

waking up to no gemini 3 was heart breaking

[zealous sparrow]

Everyone assumed gemini 3 is on nov 18

because of this

but those are just preview models

[halcyon nimbus]

grok says likely 18th nov launch too

'due to dev blogs' or something

[zealous sparrow]

likely saw this then

orr

them posting NOVEMBER 18 on the twitter community [ai insights]

because 2.5 flash and lite PREVIEWS from OLD ages are being deprecated

[hallow igloo]

Grok is the shittiest model ever

[tulip tree]

Gemini 3 before GTA 6

[night kelp]

Suggestion's:
Suggestion's channel
and new ai for 3D modeling (there are ai like that (for example the hunyuan ai that makes 3d models))

[night kelp]

W name btw

[keen beacon]

It’s gunna get surpassed in like a month or two by some Chinese model lol

[gaunt spade]

i hope so, because openai and grok arent doing good enough

claude is silent too

[keen beacon]

Bottle necked

Open AI broke need money

[gaunt spade]

only chinese models bring competion into the AI world

[keen beacon]

Yeah they catching up fast

[halcyon nimbus]

grok is good for searching stuff live, i wouldnt trust it to code or anything

[gaunt spade]

grok 4 heavy has a very weird ritual that you need to follow in order for it to make a working code for you

like you need to specify every single detail as if you're a software engineer

and it's still worse than claude 4.5 sonnet thinking

[keen beacon]

It’s just outdated lol

Grok 5 will be hyped

ChatGPT 6 will be hyped

Claude 5 will be hyped

[gaunt spade]

well Elon said it's gonna be 10% AGI

[zealous sparrow]

no it wont be

gpt 6 i doubt will be hyped

gpt 5.1 flopped

[gaunt spade]

yeah no one buys into OpenAI bs anymore

they flopped with gpt 5 and 5.1

its done

[halcyon nimbus]

6 is a lot different than a 0.1 upgrade

[keen beacon]

ChatGPT cant flop

[zealous sparrow]

5.1 bro

[keen beacon]

It’s part of the 5 series

[zealous sparrow]

5.1 flopped

[gaunt spade]

they did flop

[keen beacon]

It’s not designed in the metrics of flopping it’s more of a minor update

[spare rune]

I mean even the sloppiest model of Gemini 3 beats the highest quality of today’s model (not including Gemini 3)

So I doubt so

[zealous sparrow]

a minor update that was supposed to iprove a model

people say

it downgraded it

[keen beacon]

I wouldn’t say it to improve the model states more to make it more comfortable for a certain group of people

[gaunt spade]

to be sure, American models will not surpass Gemini 3 pro until Feb 2026

[keen beacon]

More for normies

[zealous sparrow]

fyi google owns a nuclear power plant to power ai if i had to guess

so gl to openai

[keen beacon]

ChatGPT 6 will be the model everyone is looking for

[zealous sparrow]

because openai are begging for government funding

[spare rune]

Yet still, Gemini 3 is still going to be a viable option as Gemini 2.5 is like old as time yet it’s very usable (but they nerfed it for the time being to hype Gemini 3.0)

[gaunt spade]

Google has everything lol, they own the biggest companies on Earth

[zealous sparrow]

wasnt openai begging the government for funds

[keen beacon]

No

[gaunt spade]

yes

[keen beacon]

They are asked for a co sign on loans

[zealous sparrow]

loans is getting money

[gaunt spade]

but they're a private owned company, so nobody in the government wants to fund them

[zealous sparrow]

then to pay it off

[spare rune]

Why does a multi billionaire rich ai lab company need more money anyways smth

[keen beacon]

Build out the infrastructure

[zealous sparrow]

literally this

[gaunt spade]

AI needs more datacenters

and computing power

[spare rune]

500is still a lot tho

Like a lot lot

500b*

[keen beacon]

Yeah they are asking for government help help them do sign

That way they could be able to lend more money from banks

[gaunt spade]

not against the other companies, like Google

[keen beacon]

Since they’re already a very tight credit crunch

The government will do it lol

[spare rune]

I guess

Google is like generational generation rich

[zealous sparrow]

if you cant afford to run AI instead of taking out so many loans just quit the race and let google continue..

[gaunt spade]

yeah, even if there is no AI race. Google is still winning bruh 😭

[keen beacon]

Gpogle has to be careful

Anti trust and monopoly

The government will break him up if they move too fast and too far

They’re already barely got by with chrome

[gaunt spade]

wydm

[zealous sparrow]

?

[gaunt spade]

so chrome was an illegal app?

[keen beacon]

https://apnews.com/article/google-search-antitrust-case-59114d8bf1dc4c8453c08acaa4051f14

https://www.cnn.com/2024/08/05/business/google-loses-antitrust-lawsuit-doj

[gaunt spade]

yeah but

[keen beacon]

https://www.cnbc.com/2024/08/05/google-loses-antitrust-case-over-search.html

[gaunt spade]

those are the Google search

not for Chrome

[keen beacon]

https://edition.cnn.com/2025/09/02/tech/google-antitrust-ruling-chrome-android

[zealous sparrow]

the only reason they want this

is because google is the most used search engine

[keen beacon]

https://www.politico.com/news/2025/09/02/google-dodges-a-2-5t-breakup-00540419

[zealous sparrow]

the hell is this government

[keen beacon]

Because Google is a monopoly

[gaunt spade]

yeah

[pseudo fable]

Without a retry button, there's no chance to even try; it feels incredibly frustrating.😮‍💨

[zealous sparrow]

theres one on pc if im not wrong

or maybe

your other assistant didnt gen ye

yet

[gaunt spade]

there is a retry button for me

[zealous sparrow]

Why would the government want to lose money over a search engine just being good.

[keen beacon]

Chrome is used by 3.83 billion internet users. That’s up from 3.06 billion in 2020.
Chrome has a worldwide browser market share of 67.72%.
In the US, the Chrome share of browser usage is 52.23%.

[zealous sparrow]

Also i highly doubt OpenAI will recieve gov funding as the US government is still shutdown..

[keen beacon]

Plus google search is integrated into the chrome browser

[zealous sparrow]

and how is that ilegal

I dont get their points.

[zealous sparrow]

that market share is prob from gemini.

[gaunt spade]

that's like microsoft with bing search

[pseudo fable]

The mobile webpage doesn't have a retry button; is it only available on the PC version now?

[keen beacon]

No by self

[zealous sparrow]

Has assistant 2 finished generating?

[keen beacon]

Microsoft also faced antitrust

https://en.wikipedia.org/wiki/United_States_antitrust_law

[zealous sparrow]

But google doesn't discredit other browsers...

Unlike microsoft that has edge prompt you Why dont you stay if you try to download google

[gaunt spade]

yeah microsoft wants Google to get replaced

[pseudo fable]

It has been generated, but there is no retry function key.

[zealous sparrow]

Either a bug, or it's just not there.

[amber hazel]

How to get sora 2 invite code

[gaunt spade]

not in this server

[zealous sparrow]

in the official OpenAI server theres a sora2-codes channel, people sometimes share codes there

I would recommend you look there.

[pseudo fable]

The statement "does not exist" is invalid. I could use the retry function key to try new results the day before yesterday and yesterday, but not today.

[zealous sparrow]

It's a bug then..

the LMArena CM is not online too early tho.

[gaunt spade]

anyways

[zealous sparrow]

I only can recommend waiting for pineapple [community manager] to wake up. He will check on #1343291835845578853

[keen beacon]

7FHC7H

[gaunt spade]

is gemini 3 still on canvas

[zealous sparrow]

People are saying theres limits per accounts on usage

[pseudo fable]

Sigh, well. I can only watch helplessly as the "work" that needs replacing is put to the test.😑

[zealous sparrow]

pineapple will likely be up in 6 hours.

Likely, or earlier.

[gaunt spade]

yeah, the limit is on the 2.5 pro, you get like 6-8 prompts with it

[zealous sparrow]

He will go through bugs.

[zealous sparrow]

Not that limit.

They are saying theres some kind of new limit for Gemini 3 pro canvas gens

Like you get a few per account

[gaunt spade]

idk tbh, if u use 2.5 pro on canvas you'll always get gemini 3 pro gens

[pseudo fable]

I hope my feedback will be seen after the pineapple is released. I really hope this bug can be fixed quickly.

[zealous sparrow]

I can tell your not a good english speaker, but what i meant when pineapple [community manager] wakes up

He forwards things to the team.

[zealous sparrow]

People say apparently its not the case anymore

[pseudo fable]

Haha, I don't speak English very well, it's a pity this community doesn't allow Korean.

[gaunt spade]

welp, maybe it was open for yesterday only with no limits

[keen beacon]

I doubt Gemini 3 will be a whole lot impressive

I think nano ya

Maybe some new layout for the web development and coding

[zealous sparrow]

People say gemini 3 is a router

router model

[keen beacon]

Cost saving

Ya money

Seems to be what everybody’s doing

[zealous sparrow]

I doubt riftrunners really chat specific.

It fits coding too...

In some tests, riftrunner beat canvas..

[keen beacon]

Once again, that subjective doesn’t necessarily mean it’s better

It could be just a preference that’s popular

[gaunt spade]

riftrunner also can understand your image

you can send an image to it

[keen beacon]

We got to wait till it’s in wild

Everything else is being curated to build up hype

Who knows it’s hard to separate fact from fiction since nobody really has got their hands on it like that

When it comes out, I’ll probably be really heavily rate limited

At least all the best features

[zealous sparrow]

The processing power for one gen is prob very large, so i can tell you might be right.

[whole swallow]

A man

In Thailand maybe

[gaunt spade]

lol

🐸

[quartz light]

🐊

[steel island]

i owe a camera

[quartz light]

new models: silvandra, whisperfall, tensor and 2 beluga models

many phantom models removed 🥳

phantom was poop

riftrunner still exists but likely rare

im really interested in the new ones. will try all of them out if I can

[quartz light]

tensor checked! it's an xAI model.

mercury checked (released yesterday or 2 days ago) it's a diffusion model!!!!!!!!!!!!!!!!!!!!

[split tapir]

Hello

[quartz light]

whisperfall checked! it's a kynship model.

kynship is xAI in disguise.

I just got riftrunner, it's not that rare.

[jovial hound]

6

65 فاخسا

[quartz light]

beluga checked! it's an amazon titan model.

[jovial hound]

65 TORSH

[zealous sparrow]

who is silvandra

I often get codename models when i do greentexts

[quartz light]

no idea yet

havent gotten it in like 67 prompts

[zealous sparrow]

is your prompt just identify yourself

[quartz light]

no

[zealous sparrow]

what is it

[quartz light]

uh json thingy

[quartz light]

this is "reve-v1" image model on battle

@zealous sparrow

[zealous sparrow]

its on direct too tho

[quartz light]

oh

i havent used image models in a long time

[keen beacon]

#video-arena-1 message

Yo peeps Ned a vote

Plz

Thnx g

[zealous sparrow]

banana still the goat of img edit

[keen beacon]

Damn so they got Sora 2 in here huh

[quartz light]

can i upload this here

[zealous sparrow]

you found out who silvandra is yet

[quartz light]

no

it must be really rare

this is imagen 4

[zealous sparrow]

Ima start prompting with Say what company made you
i hope i get it

[quartz light]

ye i should prob delete

[zealous sparrow]

If its rare maybe gemini 3
no im joking as-

got a beluga model dang it

[quartz light]

yeah its amazon

poop

[tawny canyon]

when will gemini 3 be released?

[zealous sparrow]

you mean silver or beluga

i think beluga you didnt mention silver yet

[quartz light]

beluga

[quartz light]

still havent gotten silvandra..

willow is still available via code arena btw

[zealous sparrow]

got ritrunner before silvandra

Maybe silvandra is codearena only? doubt it tho

@quartz light

[quartz light]

oh

[zealous sparrow]

longcat model

[quartz light]

i ran like 100 prompts

[fleet lintel]

at least we dont have to wait for riftrunner for longer now.. its releasing on Nov 18th

[quartz light]

deadass

[wintry tinsel]

Imagine a single platform where you can use Gemini 3, Genie 3, Veo 3, and Simu 2 unlimited locally this is the ideal future

[quartz light]

nice
btw riftrunner is still on arena

[wintry tinsel]

Where did you hear I’ve been checking updates hourly I even used Gemini 3 on cavas iOS a bit, it’s pretty dang good

[quartz light]

i was gonna say use opal or aistudio build but then you said local

[fleet lintel]

yeah they will do a big marketing announcement with lmsys ratings and other evals. They need raters for that

[wintry tinsel]

In a few years tech at this level should be local and snappy too, as fast as Groq

[quartz light]

i want my x28 back 🥀

[fleet lintel]

i have my sources.. 99% chance on 18 nov

[quartz light]

u mean the 17th deprecation?

[fleet lintel]

mroe reliable sources.

[wintry tinsel]

The stealth release was so weird lol, they just gave access to Gemini 3 for a limited time in some form for a day?

Only on iOS

[quartz light]

on android too

just the gemini mobile app

[buoyant violet]

Nah android also release it

[fleet lintel]

likely a wrong config that got pushed in production. happens all the time. not a big deal

[buoyant violet]

I test,same prompt with riftrunner,and gave me same answer

[zealous sparrow]

I wonder if silvandras on codearena @quartz light

longcat model so i have bare faith its good but

i want to give it a shot

[quartz light]

its not

[quartz light]

i wish models were released on friday :p

[maiden urchin]

Is Gemini 3 already in lmarena board?

[quartz light]

no

[fleet lintel]

zero chance... engineers dont want to stay over weekend to fix mishaps. only small startups can afford to do that

[halcyon nimbus]

i think flash is under rift?

[quartz light]

i know

but i just wish

[zealous sparrow]

how do we know

[fleet lintel]

we (Engineers) are not paid enough to care so much.. releasing on Friday is big NO in my team

[quartz light]

there are only a few models on code arena in general

[maiden urchin]

I heared Gemini 3 is quite good model.

Waiting for it to try.

[quartz light]

yes but perhaps in some cases it would be beneficial for companies to provide extra pay or somethin

[fleet lintel]

sometimes happens but only if there is a new contract that might go away if not acted upon soon. Releasing a new model wont fall in that category for a company of google size

[proven walrus]

yo, qq, why each gpt model 5th gen says it's from 4th gen family, but on chatgpt.com they know the right 5th gen... is there a bug or...

[quartz light]

chatgpt.com has hidden system prompt which tells them

[proven walrus]

so... we're on the 5.1, or it's capped to 4.1 even tho a 5.1 label is on it... cause it also works like a dumb brick

it's not just a small context issue

[quartz light]

no

its 5.1

it just doesnt know 5.1 released

because models only have knowledge of the past

[proven walrus]

okey... and if the knowledge is of the past, without turning on web search... is capped to the same knowledge level? (I am new to this, not trynna be a smartass) :)))

[quartz light]

thats right

👍

[proven walrus]

okey... so 5.1 just on web search on

[quartz light]

wdym

on lmarena it doesnt have tools to search AFAIK

and even if it could search

it still wouldnt know what model it is

[proven walrus]

No, let the version die

[quartz light]

wat

[proven walrus]

the gpt 5.1 is still capped to the same knowledge level as 4?

[atomic geyser]

hi

[proven walrus]

if that's so... how can I check if it's improved, to check if it's worth using as our primary API Chatbot

[quartz light]

it has more knowledge but it mightve not been trained on any ai related things

i think the cutoff is up to 2025

oh i just checked

the cutoff is up to october 2024

thats bad lol

[proven walrus]

yep./..

that's what I'm sayin'

[silent raft]

hi

how ai is doing

[keen beacon]

Takes 9 months to train a model

[silent raft]

#share-prompts

[keen beacon]

So it’s always going to be behind

[unreal dagger]

Cringe

Scammer

[halcyon nimbus]

https://tenor.com/view/lama-meme-benni-meme-lameme-benni-bois-meme-star-wars-meme-gif-18296262

[zealous sparrow]

How's your november chat, mines cold. No snow but cold.

[keen beacon]

Soon every website will cost credits lol

[zealous sparrow]

I doubt that..

[halcyon nimbus]

wasnt that the point of bitcoin, and yet literally every company just uses credits

[shadow parcel]

So sad the retry buttong gone

[tulip tree]

@halcyon nimbus

Hey I have a question

[halcyon nimbus]

?

[tulip tree]

I'm able to use deepseek r1 for free?

On open router

Infinite messages or limited?

[whole swallow]

Can I see this "exact same ui" how does it look like?

[halcyon nimbus]

free with a rate limit

[tulip tree]

What is a rate limit

Ohhhhh requests

So I can request 250 times or what

[halcyon nimbus]

50 requests per day (although if you buy 10$ of credits it bumps you to 1000 requests per day free)

[tulip tree]

What if I use Gemini pro for proxy

When I already paid for it

[whole swallow]

what's the rate limit per minute

[covert helm]

hey guys

hloo

[halcyon nimbus]

20 per minute

[halcyon nimbus]

think its the same

[tulip tree]

Ok thanks

[halcyon nimbus]

np

dang looks like ai studio only has 1 million tokens a day limit D: no wonder im running into it if true

[keen beacon]

The whole Internet is about to be filled up with slop generic everything lol

And then we’re going to train our models to reproduce the slop

I think most people avoid slop as is

[halcyon nimbus]

luckily we have anti-poision now

[keen beacon]

Wat?

[halcyon nimbus]

when training it filters out low quality slop

any model worth its salt has that these days

[halcyon nimbus]

yeah and if you get the current figure ai content WAY outweighs human content

like 90% of all text online is written by ai or something

howd we even get the storage for that?

[keen beacon]

That’s crazy

It’s never gonna stop hallucinating at this rate lol

[halcyon nimbus]

you would think with this ai boom there would be a sudden storage problem but there doesnt seem to be

synthetic data has value

[keen beacon]

It also comes with baggage

[green pasture]

hello

[keen beacon]

Sup g

[halcyon nimbus]

https://tenor.com/view/minion-hello-bello-kiss-gif-12923282

we all just killing time till gemini 3

[halcyon nimbus]

i refuse to use canvas on mobile

[gaunt spade]

why

[halcyon nimbus]

i cannot be bothered running bluestacks, and ai studio is superior and not censored anyway

[gaunt spade]

i only use it on my phone cuz Gemini 3 lol

i never had the gemini app before

[halcyon nimbus]

i type way too slow on a phone -_-

[dry nacelle]

Hey guys, is there a solution to this problem? It's been going on all day. When I start a conversation, something happens and the conversation gets deleted after I send three messages.

[keen beacon]

I wonder when the slopapcalypse is gonna happen

[halcyon nimbus]

never, that was like the y2k bug, we solved it

https://tenor.com/view/breaking-bad-gif-12280691131005323983

[dry nacelle]

It's been about 13 hours and I'm still in the same situation; I can't use the platform at all.

[halcyon nimbus]

your lucky you even have the retry button there lol

[dry nacelle]

When I press the retry button, nothing happens; everything remains the same. If I refresh the page, the conversation disappears.

[halcyon nimbus]

are you logged in?

[keen beacon]

slopapocalypse noun \ˌsläp-ə-pə-ˈkä-lə-ˌp­sē
1. A hypothetical or anticipated event in which artificial intelligence systems produce an extremely large volume of low-quality, erroneous, incoherent, or nonsensical output—including text, images, audio, video, or code—at a scale that overwhelms human capacity to filter, moderate, or utilize meaningful content.

[dry nacelle]

Yes, I'm logged in, and when I thought the problem was with my account, I tried logging in with another account, but the problem persists even after trying it from another phone.

[halcyon nimbus]

theres a 3 limit retry so its probally related to that

[dry nacelle]

Do you mean that the limit I can send messages to in the same conversation is 3?

[halcyon nimbus]

no retries are limited to 3- if you check the bug forum < i think this is being worked on by the look of it

[dry nacelle]

I hope it will be fixed quickly because I rely on it constantly for my studies and other uses.

[ember iris]

Guys need help. Am planning to get an scrollable e-invite (wedding). My idea is it should be interactive. And I want to customise the image of the bride and groom. But I can't able to make it. I don't know why. I am not a coder. But I tried something using lmarena and got one and used blogger to publish it. Please someone help me how to use the image of them... Another request which model is best to create this in website.

The one I accomplished is below
https://harini-vishnu.blogspot.com/2025/11/harini-vishnu.html

[halcyon nimbus]

use googles ai studio, way less bugs than lmarena, more daily limit too, lmarena isnt really for work/study lol =_=

but yes i hope they fix it and can up the daily limits too >.<

[tulip tree]

Deepseek r1 is so ass

[queen veldt]

istg gpt's research is living in the past

[halcyon nimbus]

thats why i still ride grok, its not good but its current

[queen veldt]

Claude's research is excellent I've had it before than i unsubscribed since soo many problems with "artifacts"

[balmy mist]

you was not saying that 8 months ago lmao

[halcyon nimbus]

it was ground breaking for like a week lol

[balmy mist]

you talking about grok right?

[halcyon nimbus]

deepseek

[balmy mist]

ohh nahh bro

[tulip tree]

Real

[balmy mist]

it impacted usa markets heavy

i would say a good month of impact on its ground breaking potential if not more

[tulip tree]

It killed chatgpt

[balmy mist]

we would not have kimi or qwen progress without it, hell i dont think some of these recent developments in closed models would have happened without deepseek idk, maybe not but deepseek truly was ground breaking

[tulip tree]

But the hype died down

[balmy mist]

yeah it did die down 100%

now kimi getting that love and even that is mid to me

[tulip tree]

[tulip tree]

Can I find that on open router

[balmy mist]

yeah

https://openrouter.ai/moonshotai/kimi-k2-thinking

its cool

but i still like 4.5 the best

[tulip tree]

Its not free

[halcyon nimbus]

256k context window zzz

[balmy mist]

no but you can use it on lmarena for free

[tulip tree]

Is there proxy

[balmy mist]

not sure, but dont worry i think g3 coming today for some reason, honestly once g3 drops what will openAI and anthropic do?

[tulip tree]

Gemini 3 gonna be flooded with requests

[balmy mist]

yeah but its free lol

[tulip tree]

But google ai pro has it better

[balmy mist]

true but why pay if its free lmaooo

[tulip tree]

The images are gonna be soooooooo goodddddddd

[balmy mist]

ik man

i cant wait!!!!

[queen veldt]

Gosh nobody can wait

[tulip tree]

But the pro model is better

[balmy mist]

imma use g3 in gemini assist, opencode everywhere

[balmy mist]

its the same model i think

[queen veldt]

I guess next Wednesday?

[tulip tree]

Nah

[balmy mist]

look now we have one 2.5 model not 2

[tulip tree]

There's 2

[balmy mist]

we dont have a pro and regular version, really????

[tulip tree]

Gemini 2.5 flash and Gemini 2.5 pro

[balmy mist]

sned the cards lmaoo

ohh thats what you mean

yeah both are going to be free bro

it always it

is*

[tulip tree]

Its gonna be Gemini 3 flash and Gemini 3 pro this time

[balmy mist]

i been using google models since they been dropping and never once paid for g2.5 pro

and i used it prob more than most people, it def will be free(like you get a certain amount of daily free usage)

and its generous at that

[tulip tree]

I only paid for the usage amount and the good image generation

And veo

[balmy mist]

ahhh

that makes sense

[tulip tree]

?av @tulip tree

[halcyon nimbus]

it WAS generious till like a month ago, skimpy as hell now

[leaden peak]

Just did a release at work 😎

[balmy mist]

really?

i havent used it in like 2 months cause i been on codex and stuff

wow

that was my end game plan damn

[cloud zinc]

gemini 3 is delayed rip

[balmy mist]

ol goofy ahh

[cloud zinc]

thats official source of delay

[balmy mist]

maybe he thinks its agi or sum and they trying to lobotomize it

[cloud zinc]

he is the cofounder of deepmind.

[halcyon nimbus]

i wish we could pay like $10 a month for more ai studio usage

[cloud zinc]

turn on billing and use ur api

[halcyon nimbus]

i rarely go over the usage but when i do its annoying

api costs way too much and random, id rather just a fixed amount

[balmy mist]

has anyone seen new benchmarks for gpt-5.1 codex?

[balmy mist]

how much is the gemini api for pro?

[halcyon nimbus]

idk, id rather just make a second google account, i doubt id use two accounts worth of usage in a day

[balmy mist]

and just make a new account and continue convo there, yeah its a bit of work but it keeps convos short and organized as a plus(so you can continue anywhere) and its free SOTA so its hard to complain for me

but i get the frusteration

[halcyon nimbus]

i reach the limit like once a month is barely worth the hassle, tbh i can just do other things, but i dont mind paying a little extra, but api calls are just insanely expensive

$10 output?? idk but thats what it says

i think i might use gemini 3 flash-lite so i avoid rate limits -_-

[placid silo]

Hi

[keen beacon]

https://news.aibase.com/news/22704

https://www.techinasia.com/news/bytedances-volcano-engine-debuts-coding-agent-at-1-3-promo-price

[halcyon nimbus]

https://tenor.com/view/silence-brand-silence-brand-dark-magician-dark-gif-23271609

[keen beacon]

https://www.volcengine.com/

5$ a month lol

[balmy mist]

not a bad deal, they need to release the api

[tired shadow]

yo that model is very good

sad that its not released yet

*released on direct chat

better than gpt 5.1 in the task I asked

gpt 5.1 (https://019a835e-489f-7809-b3e6-80f3fcf4e07c.arena.site) is too loud
the willow one (https://019a835e-489f-7f2a-9f6f-857545ff3f89.arena.site) is not loud and soft

[north cedar]

why lmarena go down every day

Something went wrong while generating the response. Please try again.

no day with no downtime

[halcyon nimbus]

https://tenor.com/view/hang-first-time-smiles-gif-14991372

[north cedar]

100 time

[balmy mist]

willow, who's model is that?

[zealous sparrow]

willow is early gpt 5.1

fyi

[neat apex]

Does mercury comming back to Lmarena means anything? xd

[keen beacon]

https://youtu.be/HxoR1GkiS0c?si=Adk2q00g1i3mEACe

[gilded trail]

https://www.rival.tips/over-the-years?provider=openai&challenge=svg-layout

[queen veldt]

This is cool

[shell drum]

help me gng

interaction failed while generating vid

[frozen swift]

no more retry button? or it's just me?

[shell drum]

does the prompt part being higher than 2000 characters or like whatever the discord regular limit will it still work

[hollow copper]

hey everyone!

I am looking for a full stack dev to develop and create an MVP demo for my business.

im trying to create an internal chatbot for Tier 1 engineers in IT companies, that can help them by accessing knowledge bases, suggest auto resolve statements, and provide real time assistance.

please let me know if there is anyone out there looking to work with me on this project!

[empty stump]

Gemini 3 when

[robust sluice]

they removed it

[keen beacon]

man this reminds me of ChatGPT 5

All the hype the delay

[cloud zinc]

who is hyping?

[keen beacon]

https://tenor.com/view/bruh-gif-17563483835665589874

[frozen swift]

thx, but why?

[keen beacon]

https://youtu.be/J1mA8Bkq0xs?si=yNd3PH-QifzTB5jr

😂😂😂

[cloud zinc]

https://tenor.com/view/gold-medal-gold-win-winner-you-did-it-gif-15707454293895969047

[keen beacon]

Everything is gunna be on credit

[cloud zinc]

good

[fresh fulcrum]

guys lmarena is this really better than all the others

[keen beacon]

https://tenor.com/view/what-are-you-trying-to-say-benjamin-engvid-what-do-you-mean-wdym-gif-24543534

[halcyon nimbus]

needs a comma and a questionmark

[robust sluice]

you can check in feedback room, they talk about it

[halcyon nimbus]

https://tenor.com/view/this-is-all-i-got-man-this-is-all-i-got-desperate-come-on-ken-jeong-gif-11868058

i only use lmarena, ai studio and grok

[keen beacon]

Like ai pictures they r cool videos in ai

Just what I needed

[lyric oxide]

https://youtu.be/JQ4pkO4QFLg?si=aQnU24pWxEvbXM4L

[keen beacon]

Where the chains necessary?

I like the details the Nike shoes on the orca

Bootleg dbz

🤣🤣

Mini max is so wild

[cloud zinc]

https://tenor.com/view/side-eye-gif-17856173530980245283

[keen beacon]

They having a horror contest

But I’m not really into that stuff

Normally this prompt generates Dragon Ball Z stuff, but for some reason not in mini max

[hoary elbow]

Since LM Arena put the highest ranked models on the top of the direct shot selection and qwen 3 max Is in the top, how good is it

Men

Nvm

AutoCorrect, I actually scrolled up and it turns out 2.5 pros at the top. I’m sorry.

[cloud zinc]

2.5 pro is at top

[hoary elbow]

Yeah, I just clarified that

[hoary elbow]

Did they make 2.5 Pro better because it’s at the top

[kind oxide]

when is gemini 3 coming out does someone know

[burnt sinew]

November 18

[cloud zinc]

december

[burnt sinew]

I dont think we can trust them can we?

[burnt sinew]

Well someone else at Google said the 3.0 pro preview was fake when it was real

[cloud zinc]

cofounder of google deepmind

[burnt sinew]

Okay?

[hoary elbow]

Maybe it is coming out in December maybe it said 11 for the model because it was a testing model maybe

[cloud zinc]

u are not gonna trust the head of ai company releasing it?

[hoary elbow]

And also, it’s a developer

Of Gemini, so we probably need to trust him

[burnt sinew]

Someone else also at Google lied about gemini 3 stuff

[cloud zinc]

someone who? dont fall for random leaker

when the official cofounder of deepmind is saying it, its true

[burnt sinew]

Let me find it i forgot tbh

[hollow crypt]

depends on what he means Gemini 3 generally or for Sima 2
Its very weird answer either way

[hot ember]

hi

[lyric oxide]

it's not

[fiery gull]

I cant send like or dislike (my life is nosense now) @echo aurora

[cloud zinc]

he means "see gemini 3" [for public]

[hollow crypt]

even if so depends on what a bit longer means, days , 1 week, more ?

really up to interpretation

[cloud zinc]

"a bit longer" thats more than 1 week. 2-3 weeks

[hollow crypt]

highly highly doubt , more & more people are getting access to it + its already in IOs

[burnt sinew]

Logan

[whole swallow]

Ativar o Windows

[fiery gull]

I don't want turn microsoft more billionarie 😠

[cloud zinc]

who that

[whole swallow]

U brazilian?

[burnt sinew]

He works at Google

[keen beacon]

Makes sense

[burnt sinew]

https://x.com/officiallogank?lang=en

Logan Kilpatrick (@OfficialLoganK) on X

Lead product for @GoogleAIStudio + the Gemini API. My views!

X

[keen beacon]

Before after

[cloud zinc]

what did he lied about

[keen beacon]

Making sound like they got something special for us

[upbeat dune]

Who's waiting for gemini 3.0?

He must become the best ai

[burnt sinew]

@fixbymeasure said that there was gemini 3 preview available and Logan said it was fake, when it wasn't

[keen beacon]

Testing different variance

[fiery gull]

yep

[keen beacon]

There’s probably like 13 versions of Gemini3

[burnt sinew]

https://x.com/fixbymeasure/status/1986464440586211707?s=20 @cloud zinc

simon (@fixbymeasure)

gemini 3.0 pro is now (somewhat) available on google vertex to a few, limited regions and accounts as gemini-3-pro-preview-11-2025

rn, you can use it via some chinese api aggregators:

- https://t.co/spbgSWGHvz (limited free credits available)
- https://t.co/K6BXpF5YFX

there

X · Nov 6, 2025, 4:03 PM

[burnt sinew]

Yeah there's a lot so far

[echo aurora]

Hmm are others seeing this too?

[fiery gull]

I'm curious too

[keen beacon]

They’re probably just trying to figure out which model is the most compatible

[fiery gull]

I'll try in my smartphone

[keen beacon]

It has to do better than the open source models where it counts

[fiery gull]

ops, this problem is just my pc

[fiery gull]

in the next time I'll check this ok

[cloud zinc]

open source are not multimodal

[keen beacon]

Not everybody’s looking for multimodality

If I had to guess I’d say that most people don’t utilize multi modality. I think people are more interest oriented.

https://youtu.be/vPaF6TypfX8?si=SZdZV1u9RM_PH7ur

Oh geee

I better go make myself safe

[soft matrix]

AAAAA RETRY FAIL

[quick quail]

Hello guys! We all gud yh!

[viscid timber]

yo what if rain-drop was gpt-5.1

[balmy mist]

gpt5.1 codex is cheeks everywhere besides codex lol

[gleaming roost]

🤔

[balmy mist]

ur linkedin post?

[surreal creek]

that's me bruh

[fleet lintel]

stop spamming. there should be some social etiquette test before people are allowed on internet

[surreal creek]

ban him please

this is just self-promo slop

go away genuinely engagement baiting is not welcome

[queen veldt]

You joined here just to promo...

[drifting crow]

I wanted to see his promo

[keen beacon]

Mid journey is free on meta ai currently

And videos

[main nimbus]

Gemini 3, am I right fellas?

[keen beacon]

Mid journey vids

[next thistle]

Give me Arena logo

[gleaming roost]

👀

[keen beacon]

Free right now on meta.ai

Mid journey image plus video generator

[next thistle]

Meta AI lOlzZ
Now it's not works good.

previous version was good

[echo crane]

Hye guys i have a question,can i ask it here?

[keen beacon]

Ya kinda laggy

[next thistle]

Any video generator AI for 10 second?

[cloud zinc]

this sora 2?

[keen beacon]

No meta ai

Mid journey

[cloud zinc]

can it make 10 second video?

[echo crane]

why is it that in the direct chat in the lmarena when i select claude 4 opus,and ask what model r u,it says 3.5 sonnet? is the model name all a fake? similary with gpt 5 it said iam from the family of gpt 4

[keen beacon]

It can extend

[cloud zinc]

yes fake

[next thistle]

Mid Juarney only for rich peoples 😂

[upbeat dune]

guys, what is better? 4.5 sonnet or 4.1 opus

[echo crane]

4.1 opus ofc but varies by your need

[cloud zinc]

looks awful

[echo crane]

bruh

[keen beacon]

Ya

Let me try more complex

[upbeat dune]

for coding

[next thistle]

Bike stunts

[echo crane]

claude 4.1 opus is built for complex coding

does this lmarena chat models have a rate limits or something or is it basically unlimited

[cloud zinc]

bike stunt

[next thistle]

Try to make epic war lOlzZ

[keen beacon]

This is the same prompt I used here, but I don’t know how to use mid journey really well

Quite a difference

[next thistle]

What happens with back tire?

[keen beacon]

Oh, might as well show you the sora version also

[cloud zinc]

why cartoon?

[keen beacon]

They all come out different

[next thistle]

15 second sora?? Really??

[cloud zinc]

what

[keen beacon]

Ya

[next thistle]

Awesome!!

[cloud zinc]

sora is 25 seconds

[cloud zinc]

its not moving?

[keen beacon]

Yes

[cloud zinc]

wow this midjourney?

so good

[keen beacon]

No Sora

[cloud zinc]

so much better

[keen beacon]

Yes

[next thistle]

If we generate very detailed prompts related videos under 5 sec then 15 sec is enough 😅

[next thistle]

Hehhe... Rocky Balbua

[keen beacon]

Some dude I shared this account with

I was making some really awesome anime. I’ll show you some.

[cloud zinc]

shared account?

[next thistle]

We call it "Jugaad" lOlzZ

Sora also paid version

Not free

[cloud zinc]

what is jugaad

[cloud zinc]

this midjourney?

damn

😲

[keen beacon]

No it’s what he was making

[cloud zinc]

thats sora 2 pro?

[keen beacon]

I think so

[next thistle]

Collecting fog from here and there, and showing a great work.

[keen beacon]

I got some good boxing stuff too needs edits

[next thistle]

AI not generate hitting sequence.

[keen beacon]

One se.

Sora lags so hard

[next thistle]

Fight together shot will not generate

[next thistle]

So how it works?

[keen beacon]

1 sec

I was tweaking some prompts

Difference in language really affects how it comes out

Seems to be the most dramatic

Is Japanese

[cloud zinc]

whats this at the end

[keen beacon]

Yeah, these last shots they hallucinate the most

Like the fast-paced stuff

And then you lose your characters for some reason

It’s probably leaning towards the training data

[solid brook]

@echo aurora Can you guys add gpt 5.1 high?

Rn it does not reason at all

[cloud zinc]

gpt 5.0 high much better than gpt 5.1

[sharp mirage]

why ?

to now i see gpt 5.1 is better than 5.0 high

[cloud zinc]

they use different budgeting system for 5.1

[sharp mirage]

so?

[stray aspen]

Explains math better

And actually knows how to render damn latex code

[sharp mirage]

ah :

[sharp mirage]

but it better at understanding and better powershell codes

[knotty fable]

Latex code? Is that something NFSW-ish? 😸

[stray aspen]

No

It's this

Basically rendered math equations

[knotty fable]

Sorry for joking.

Lets chill with a:

[knotty fable]

Hmm, someone in room 2 made an ad in Swedish and actually managed the AI to speak in the language.
The AI did remarkably fine until the very last word - when it failed badly.

[unborn nebula]

Is it true that the code arena can only create a single file?

[cloud zinc]

yes

[unborn nebula]

That is unfortunate. Is it possible for them to develop it into a full-stack?

[sharp mirage]

anyone having any idea about any good ai that good at js and hrml

html

[gaunt spade]

Gemini 3 lol

cuz its far better than any model

at coding

[sharp mirage]

There is no Gemini 3

/:

[sharp mirage]

It comes?

[gaunt spade]

u can use it on gemini app, using canvas

[sharp mirage]

Hmmm

Okay

[gaunt spade]

it doesnt show the model

[sharp mirage]

Can you send url ?

[sharp mirage]

:/

[gaunt spade]

but u know its gemini 3

cuz its better at coding

[cloud zinc]

they removed it?

[gaunt spade]

u need the mobile app

[gaunt spade]

did they

imma test it rq

[sharp mirage]

Okay go on

[gaunt spade]

yeah imma test it and see

if its still on

[gaunt spade]

no i guess, but it got nerfed so much now

[cloud zinc]

its using 2.5 pro now

[gaunt spade]

no 2.5 pro is MUCH MUCH worse

[cloud zinc]

i have confirmed.

[gaunt spade]

its a very nerfed gemini 3

maybe gemini 3 flash

[cloud zinc]

nah its gemini 2.5 pro.

[gaunt spade]

still gives a little better results than gemini 2.5 pro on AI Studio

[cloud zinc]

gemini 3 flash wont be better than gemini 2.5 pro.

[gaunt spade]

maybe it will

[cloud zinc]

it wont

[gaunt spade]

with thinking on

[cloud zinc]

nope

[gaunt spade]

its two models only, Gemini 3 pro and Gemini 3 flash

welp i think

we gotta stick to riftrunner now

[cloud zinc]

flash-lite

[gaunt spade]

its the only Gemini 3 checkpoint now available

back to lmarena lol

also there are tons of new hidden models now

added to lmarena

@quartz light yo

are the new models any good

or just riftrunner

neon and jaguar 🤔

new models added to lmarena

[quartz light]

oh thx

ill check em

but those are 1 word codenames

usually lame

like 5.1

what

so its qwen

yes its qwen

i got it again

the ratelimit on switching chats is crazy

[sharp mirage]

How can I send about 20k layers to lmarena?

I sent it by txt to glm 4.6

Who

twenty thousand letters

Oh

K

My English is not good 🙂

[quartz light]

not much

its a problem with chat switching lol

btw

bridge-mind is still alive yall ❤️‍🩹

aka gemini 3

[quartz light]

i just got jaguar

finally

it says its mistral :p

bruh

just got neon too

its mistral too

wait what

i just got bridge-mind 4 times in a row or somethin

and it didnt error like it usually does

oooo

DUDE

I JUST GOT IT AGAIN

WHAT

ok the streak broke but i got it so many times after that

dude

its so common now

lets go

@burnt sinew @gaunt spade

[balmy mist]

this in the arena?

[echo sinew]

@lofty eagle Please, read our guide in ⁠⁠https://discord.com/channels/1340554757349179412/1397655624103493813 to learn how to properly prompt the bot.

[balmy mist]

2.5 pro is that far from riftrunner like at all

very close

[winged mortar]

HELLO

[tardy bolt]

anyone knows how to add sound effects to the video

with ai

[echo sinew]

Hi. Sorry to say that in the video arenas, getting a model that adds audio/sound to your videos will be completely random. You have to try your luck.

[keen beacon]

You can try to add specific prompts to sound

But like he said it’s random

I typically don’t like any sound in my video, but I always add sfx effects in my prompt

"High ambient light",soundscape:"Loud, energetic shindō (vibration/sound effects), creaking mechanical sounds, strong wind effects"} /

Not gunna lie some yall be making some questionable content lol

Borderline I’ll watch some of the videos in the arena from time to time

Look at this prompting lmao

“The girl on the right leans in to smell the girl on the left, then smiles. Afterwards, the girl on the left smells the arm of the girl on the right. Both girls begin laughing”

🤣🤣🤣

Like cmon bro can u be anymore obvious.

[ocean ferry]

the lord is teasing us too😭

first Logan then Sundar?😭

[keen beacon]

69%

Hmm

Well, we do have two pieces of evidence that make no sense

The original

Updated

Unless they’re extending the lifespan this could also make sense

Or it could be to add some more smoke into the mix knowing damn well it’s gonna be out on the 18th

Hard to say

December marks the end of the fourth quarter (Q4) and the close of the business year for many companies. Releasing major products or updates at this time can align with internal planning cycles, end-of-year budgeting, and the start of planning for the next year (Q1).

https://tenor.com/view/aj-holmes-tcb-tin-can-bros-idk-i-dont-know-gif-21538417

[ocean ferry]

ohh shitt no pls Sundar no

[keen beacon]

We have to pull out the calendar to look at our analysis further

The week after next would be thanksgiving a 3 day work week almost

Would make no sense

[lime coral]

18/3=6/2=3. Gemini 3 dropping the 18th. CQFD

[keen beacon]

Next week would be the ideal time

17-19th

Maybe 20th max

Cause we know they’re not gonna drop it on a Friday

Then we have this guy

But we’re getting mixed messages here cause he has a 😊

Does that mean they got something big planned for us or is he teasing us? confusing words “a bit longer.”

https://tenor.com/view/inspector-gadget-chin-scratch-hmm-thinking-gif-16686907400137345357

We have to turn to our last piece of evidence