#general

1 messages Β· Page 333 of 1

silent tree
#

without authorization

#

lol

#

hes using banana against me

#

banana isn't even a mod

storm dust
#

screenshotted that watermelon

median smelt
#

yea

silent tree
#

why pressure on him

median smelt
#

exactly

sullen sable
#

I touch grass btw

median smelt
surreal zephyr
#

@light sleet

storm dust
silent tree
#

no point

sly cedar
silent tree
#

who are u gonna send this to

urban herald
#

@light sleet

silent tree
#

the mods

median smelt
#

SCREENSHOTTED BOOIIIIII

#

GET SCREENSHOTTEDDDD

#

LIL BRO

median smelt
#

OMGGG

light sleet
#

hello guys

median smelt
#

banana

silent tree
#

wsp

storm dust
silent tree
#

🍌

sullen sable
#

A screenshot of someone saying "Screenshotted that watermelon"

storm dust
#

im not leaving the server or uninstalling the extension until banana says so

#

i dont know what to do

#

so im doing nothing

sly cedar
median smelt
#

@storm dust Threatening to report me to the mods because I called him out for pentesting unauthorized

storm dust
#

im there with no purpose anymore

median smelt
#

@light sleet

obtuse heart
#

what the hell is going on in this chat 😭 ..

silent tree
#

@surreal zephyr ALTMAN MOGGED HIM

storm dust
obtuse heart
#

buncha kids arguing abt some lm arena extension

sullen sable
#

Hi lol (I'm 6'10, I like matcha, and I wanna end period cramps)

storm dust
sly cedar
median smelt
#

in a little bit of course I hope im in dont call that song

obtuse heart
#

take it up to pineapple then, why r yall arguing in the public chat

storm dust
#

i didnt know that but it was obvious

storm dust
median smelt
#

everyone SHUT

silent tree
storm dust
#

i even dmed to modmail

surreal zephyr
sullen sable
#

Son😭

median smelt
#

theyre gonna ban you for pentesting

silent tree
storm dust
surreal zephyr
silent tree
median smelt
#

spud yesterday

surreal zephyr
silent tree
#

spud 2 days ago

#

😈

surreal zephyr
sly cedar
#

Blud gpt 5.5 πŸ€‘

surreal zephyr
silent tree
sullen sable
#

Spud potatoπŸ˜‹πŸ˜‹

obtuse heart
#

mythos release

sly cedar
median smelt
#

Bro 5.5 was released tomorrow

surreal zephyr
median smelt
storm dust
#

hey banana

light sleet
#

ampro is sam altman undercover btw (hes pretending to be a member)

sly cedar
storm dust
#

i know you got pinged a lot

median smelt
#

BM MASS IGNORING

HAPPENING: RIGHT NOW

storm dust
#

im sorry you were involved

median smelt
#

EVERYONE IGNORE BM

urban herald
storm dust
median smelt
sullen sable
#

damn

sly cedar
median smelt
#

Pentester unauthorized

urban herald
median smelt
#

you just admitted to committing a crime on a chatting app

urban herald
median smelt
#

you admitted tho

#

Thats the proof

urban herald
median smelt
#

fair

urban herald
#

yes

median smelt
surreal zephyr
#

πŸ€”

median smelt
#

I am sneakers o goat

sullen sable
surreal zephyr
#

good person

#

67

median smelt
sly cedar
sullen sable
sly cedar
light sleet
median smelt
obtuse heart
#

this chat is the definition of in heaven looking at hell

surreal zephyr
sullen sable
#

yes, chill

surreal zephyr
#

actually im cancelling spud

median smelt
#

Tralalero tralala says a blasphemy in the original video

surreal zephyr
#

lelele

median smelt
#

Spud also did

surreal zephyr
#

@gork sam saltman eating spud

#

πŸ₯€

sullen sable
#

I touched dried grass

sly cedar
sly cedar
median smelt
#

yea

#

That ones worse tho

sullen sable
median smelt
#

67

sly cedar
sullen sable
tulip kindle
#

When will Gemini 3.1 Pro return?

sullen sable
#

I'm crine

sly cedar
urban herald
#

who wanna unauthorised pentest with me

median smelt
#

no

light sleet
light sleet
#

I'm still fan

#

Release gpt 67

#

Plssss

#

U promised

sullen sable
sly cedar
#

Not again

median smelt
#

Bro i just made something so good wait

sly cedar
#

Im getting brainrot πŸ˜­πŸ€‘πŸ‘οΈπŸ”₯πŸ˜‹πŸ§‘πŸ»β€πŸ¦²πŸ™πŸ»

median smelt
surreal zephyr
median smelt
#

spudalero spudala

sly cedar
surreal zephyr
sullen sable
#

Do u like spud potato

urban herald
median smelt
#

guys

#

We are all using websim bot

#

Trust

sly cedar
sullen sable
#

Smev, do u eat the fish

sly cedar
sullen sable
#

Do u like coal

sullen sable
#

Son 😭

sly cedar
sullen sable
#

wuthering wave characters or genshin characters

urban herald
#

πŸ₯€

sullen sable
#

I need to lock in 😭

sly cedar
sullen sable
#

Bro 😭😭

median smelt
#

Bradar

sly cedar
median smelt
sullen sable
#

I think I'm straight guys

urban herald
median smelt
#

no way

sullen sable
surreal zephyr
#

guys maybe yall shouldn't

sullen sable
#

Who wanna play

urban herald
surreal zephyr
#

if mod 9, 10, or 11 wakes up yall are dead 😭

urban herald
sly cedar
surreal zephyr
#

8, 5, 16, my bad

surreal zephyr
#

LMAO thats the exact reason i got muted

sullen sable
#

Might as well take the risk

surreal zephyr
sullen sable
#

We only get to live once

surreal zephyr
#

🀣

sullen sable
surreal zephyr
#

we live in 1984

sly cedar
sullen sable
sly cedar
#

Bradar vat is this, delet roblos πŸ”₯

sullen sable
#

I'm 95 yrs old

urban herald
surreal zephyr
#

dang grok is eons behind sora

sly cedar
#

"We"

urban herald
#

@daring rock ur a bot

sullen sable
surreal zephyr
#

bro is gona get executed 😭

median smelt
#

What is happening

sullen sable
urban herald
sly cedar
#

You just leaked my weak source

urban herald
surreal zephyr
#

no joking in big arena πŸ₯€

sullen sable
sly cedar
urban herald
#

cats and dogs

sly cedar
#

But i have a dog named cupcake

sullen sable
#

do u like bacterias

urban herald
surreal zephyr
#

Sam

sly cedar
surreal zephyr
surreal zephyr
sullen sable
#

where r u guys from, I wanna touch grass there

sly cedar
median smelt
median smelt
surreal zephyr
median smelt
#

Bro

sly cedar
median smelt
#

Thats not triple t

surreal zephyr
#

Can we add 67 to global slur list and blacklist it

surreal zephyr
#

After he lost it all

sly cedar
sullen sable
#

Is that a cat

urban herald
sly cedar
#

Why

sullen sable
#

Is that ur kryptonite bro

sly cedar
#

STOP

#

This is broke my heart

sullen sable
median smelt
surreal zephyr
#

scrolling thru my gif folder knowing one missclick would get my ahh banned 😭

#

(joke mods dont ban me)

median smelt
#

sora 2 eons away from seedance 2.0

surreal zephyr
median smelt
#

im crine

surreal zephyr
#

bro i cant find it

#

anywhere

#

fake ass

sullen sable
#

Once upon a time

median smelt
#

Bro

surreal zephyr
#

bro

median smelt
#

use Websim bot

surreal zephyr
#

theres no seedance website

median smelt
#

because its the best bot

surreal zephyr
#

whats that

median smelt
#

it has free seedance

surreal zephyr
#

where

median smelt
#

and sora

surreal zephyr
#

give

median smelt
#

dms

surreal zephyr
#

πŸ™

median smelt
#

wait

surreal zephyr
#

give

sullen sable
#

😭

surreal zephyr
#

😭

sly cedar
#

Ohio warn πŸ™πŸ»

#

What i even do bro 😭

sullen sable
storm dust
sly cedar
sullen sable
#

What

sly cedar
#

Whats the algorithm

storm dust
#

is it ok if i ask you something?

sullen sable
sly cedar
sullen sable
storm dust
#

could you please check my bug reports people claim that i am "pentesting"

#

i need your opinion

sullen sable
#

Wait

median smelt
sullen sable
#

uolol

surreal zephyr
#

whats a pen

sullen sable
#

penetration ig

#

idk

#

ballpoint pen

#

To write

median smelt
#

Basically

sullen sable
#

Are you chinese

median smelt
#

no

storm dust
median smelt
#

because im sure you did

#

now zippy that mouthy shutty

storm dust
#

im blocking you

median smelt
#

omg u blocked me im so sad

#

Im crying now

#

Look at what you did ohlol you offended me

surreal zephyr
#

seedance is not even 1/10 as good as sora

median smelt
#

Okay anyway

storm dust
median smelt
#

ur kinda right tho

storm dust
#

he's a ragebaiter

echo aurora
#

Okay not sure why people are fighting with each other, but going to timeout everyone involved

median smelt
#

okay

#

see yall soon

storm dust
#

please check my bug reports they are important

median smelt
#

you pentested unauthorized mate

#

Anyway

#

Everyone zip it

surreal zephyr
#

🀣

#

anyway

#

when spud

#

πŸ™

#

(im not involved)

#

i miss sora

urban herald
surreal zephyr
#

sora actually looked realistic unlike all other

sullen sable
#

Oh damnn

urban herald
#

NOO

#

TIMED OUT

#

😭

surreal zephyr
#

dead

honest verge
#

What's the difference between GitHub coplot and Microsoft copilot

golden ocean
#

microsoft copilot is just chatgpt and is bad

#

github copilot lets u choose models like claude or somehting, its not a model itself

#

i think

light sleet
#

Which u prefer

trail oak
#

do u guys have agent code

#

Mode

light sleet
#

some do

#

some dont

trail oak
#

oh ok

rocky geyser
#

Hii

rocky geyser
trail oak
#

nice, I don't have tho

rocky geyser
#

Oh okay

jaunty talon
dull mulch
#

Hi, can someone please explain why I can't use Claude Opur 4-7 thinking in text arena when I can see it in leaderboard?

light sleet
light sleet
#

βœ… βœ… βœ… βœ… βœ…

rocky geyser
pseudo sage
#

why my arena.ai not working or answering any ai

#

Something went wrong while generating the response. Please try again.
Trace ID: bb2e676a-dfef

surreal zephyr
vale quest
#

Fr

spring oar
#

in llm arena this is without thinking ?

vale quest
echo aurora
vale quest
#

Never gives a crap to their users except for the data they earn money from when they vote models

echo aurora
#

It's in Battle mode

spring oar
# echo aurora Yeah

So the model without the β€œthinking” feature is more powerful than models that do think, like the Gemini 3.1 Pro

vale quest
#

Yo users are dropping massively because of it

echo aurora
vale quest
echo aurora
#

?

surreal zephyr
echo aurora
sterile tartan
echo dome
sterile tartan
light sleet
#

Popus*

sterile tartan
#

Is Slopus 4.7 on Web/App

sterile tartan
echo dome
stray aspen
#

c@light sleetare you the pineapple

indigo knoll
#

Will Opus 4.7 be added to direct and side-by-side?

stray aspen
sterile tartan
sterile tartan
obtuse sequoia
indigo knoll
sterile tartan
light sleet
#

@echo aurora For me, all models are showing the error

#

Something Went wrong

still musk
#

Maybe tomorrow

light sleet
#

Gpt shows it too

obtuse sequoia
#

doesn't work

jade sky
#

seems like gpt gemini and claude are down atm

marsh schooner
echo aurora
echo aurora
rigid holly
#

opus 4.7 reaches number 1 on the list

can not even use opus 4.7 as opus models got nuked off the site

gray vortex
echo aurora
echo aurora
obtuse sequoia
spring oar
surreal zephyr
echo aurora
sterile tartan
surreal zephyr
#

6 spud

sterile tartan
#

No need of Slopus and Myth OS

sterile tartan
light sleet
#

they gonna release 5.5

surreal zephyr
#

Native voice and video is peak

light sleet
#

watch

surreal zephyr
surreal zephyr
light sleet
#

v2

sterile tartan
light sleet
#

tomorrow πŸ˜”

surreal zephyr
#

Theres no 5.5 😭 😭

sterile tartan
light sleet
surreal zephyr
#

Images v2 might be part of spud

light sleet
#

I think images v2 tomorrow

surreal zephyr
#

Then we would be cooked

#

Spud 6.0 is

#

Omni, new base

light sleet
#

GTA 6.0

#

Gta 6.0spud

surreal zephyr
#

Yes

#

Cuz 5.0 isnt multimodal base

#

6o is

sterile tartan
#

Spud will Mog All

obtuse sequoia
#

@minor shoal

#

πŸ‘‹πŸ»

outer flicker
#

gork 4.3 in beta guy

light sleet
echo aurora
#

How do you manually update in? Just go back to Code Arena and tell it to add it?

light sleet
#

yea xd

echo aurora
#

Lol nice

light sleet
#

this time for some reason i couldn't copy the text so I wrote it manually too

#

as soon as the notification popped up

#

Asked glm to create new article lol

echo aurora
#

I wonder if there is a way to update it automatically

vernal raft
#

@echo aurora Just a quick question.

Ofc there are words that are blocked from the input prompt (I mean the tos), but what if the model itself outputs those blocked words? Do I still see the output?

stable ridge
#

i could not able to see claude opus 4.6 and 4.6 thinking in my agent from last 2 weeks
did anyone else also facing the same problem

echo aurora
# vernal raft <@283397944160550928> Just a quick question. Ofc there are words that are block...

Ofc there are words that are blocked from the input prompt
My understanding is it'll take into account context as well, compared to specific words it doesn't like (small difference but wanted to point out incase helpful information).

what if the model itself outputs those blocked words? Do I still see the output?
I don't believe so as there is a check on the input and output. At least there is for Image & Video, I assume same happens for Text

echo aurora
surreal zephyr
#

No its all 4.0 base

#

First digit is base

#

Yooo kimi k2.6

half veldt
#

gpt image v2 is out

light sleet
#

Very real

#

very real Image v2 is out! 1!11!1!

#

no cap 100% real skibidi edition

half veldt
#

im not joking

light sleet
#

in my dms

half veldt
#

check twitter lol

#

its rolling out to all regions currently

light sleet
half veldt
whole sundial
light sleet
#

banana disapproved

#

By banana disapproved I meanf

#

He disapproved me

#

Images v2 is indeed rolling out

light sleet
whole sundial
#

yeah i'm still getting image 1.5 though

slender thistle
#

V2 is not out

whole sundial
light sleet
formal snow
#

Gemini 3.1 and Claude 4.6 Opus are gone from Direct.

light sleet
harsh blaze
#

Pineapple . I think each chat should have an token limiter like exact number to know how long we need to know that it gonna end.

light sleet
#

Pineapple what's the first thing youre gonna generate when gpt image 2 is out

#

I'll generate Sam Altman holding a pineapple

formal snow
proud bobcat
#

Kimi K2.6 just released

#

This is really good

gray lion
cedar citrus
#

yoo

#

kimi 2.6

formal snow
cedar citrus
gray lion
#

oh yea kimi 2.6 is here

surreal zephyr
#

Qwen 3.6 max is out too

cedar citrus
#

yo

gloomy onyx
cedar citrus
#

finally

sterile tartan
surreal zephyr
#

5.4 pro xhigh solos

sterile tartan
sterile tartan
light sleet
surreal zephyr
#

spud is going to be really fast btw

#

like fast fast fast

#

and smart as hell

#

new architecture

civic plaza
#

I saw Opus in the announcement and thought it finally came backπŸ₯²πŸ˜‚

surreal zephyr
#

v2 or v1.5 πŸ€”

#

but damn it was fast

wicked sage
#

hi guys

echo aurora
sterile tartan
echo aurora
wicked sage
errant sand
#

Kimi 2.6 came out

wicked sage
#

my goal is to.. get into lmarena leaderboards
EVEN if its like top 200000

errant sand
#

Not sure how good it is

surreal zephyr
#

v 1.5 takes around 3-5 minutes

wicked sage
#

pineapple can you give me a h200 so i can train my ai ok thx /satire

surreal zephyr
#

i think this might be 1.5

#

but its suspiciously fast

sterile tartan
#

It's 1.5

#

Try Dario Amodei

echo aurora
sterile tartan
#

If it gets it it's 2

echo aurora
errant sand
surreal zephyr
#

old youtube?

#

definitely 1.5

#

but 1.5 mightve improved a bit since last i checked, bc it used to be much worse than nano2

median smelt
#

was I not timed out

wild ridge
#

Does anyone know which AI this is? It created me a beautiful perfect website. I cant describe how good it is (Its just for learning for a friend its all in german)

surreal zephyr
wicked sage
#

but ill try my best

wild ridge
wicked sage
#

it mightve have been better to just finetune a model but f it we ball

wild ridge
#

I wanted the simple desgin

sharp iron
#

i keep getting Something went wrong with this response, please try again., no matter the model i try

wild ridge
#

and its just too perfect

surreal zephyr
sharp iron
#

i've tried quite abit

wild ridge
surreal zephyr
wild ridge
#

I think that will be a crazy ai model

wild ridge
median smelt
desert pendant
#

-# jk

wicked sage
#

oh go for it lmao

desert pendant
#

good luck friend

wicked sage
#

ty ty

desert pendant
#

hope you get atleast in top 4

#

or 3

surreal zephyr
#

yup 1.5, doesnt get abrams x correctly

wicked sage
#

maybe top 100

desert pendant
wild ridge
#

Ive already created lots of websites with claude and so on but something is diffrent on this one.. I cant explain it..

#

even if its just that simple.. something feels 1000 times better than on other ai websites

half veldt
#

whats the best model except claude models for frontend design

sharp iron
#

codex 5.3

#

imo

storm dust
#

oh

#

thanks pineapple

#

or whoever removed the time out from me

sharp iron
#

nw done

pseudo hemlock
#

Anyone know if opus 4.7 is in search battle mode?

#

Or how it compares to opus 4.6 search

wicked sage
#

opus is removed due to high cost

rocky geyser
pseudo hemlock
wicked sage
#

oh wait im so stupid

#

isnt there a search mode in leaderboards

rocky geyser
#

yea

sharp iron
#

im a patient guy tho so i just tweak him to what i want properly

#

100% takes abit tho

rocky geyser
#

Yea i guess

#

Ive been loving sonnet 4.6 its an amazing model!

echo aurora
pseudo hemlock
#

Dang ok ty pineapple

wicked sage
#

if you even have one that is

rocky geyser
sharp iron
#

is sonnet good, ive been using codex mainly cause i tried sonnet once, token limit reached πŸ’”

rocky geyser
#

Even tho i have always disliked sonnet models until now

echo aurora
rocky geyser
sharp iron
echo aurora
austere vine
#

Hey folks at the Large Model Arena, when are you planning to add a ranking for model vision capability tests?

acoustic garden
#

Why did the Opus models disappear from the arena?

rocky geyser
echo aurora
austere vine
light sleet
#

Not as good as duct tape

#

Seedance 2 fast

storm dust
#

that looks good

light sleet
#

Ty

echo aurora
echo aurora
vast fern
#

opus 4.7 when

echo aurora
wicked sage
#

idk, sterling

still musk
#

GPT IMAGE 2 IS OUT OMG

light sleet
#

yea but it's releasing by region

still musk
#

NOW

#

😭😭

light sleet
#

@echo aurora U heard the man, its time to ask the team!

remote ocean
#

hello @echo aurora i wanna ask why opus is no longer available in lmarena?

still musk
# still musk

If it has been released in some regions, please release it now via direct chat in Arena AIπŸ™πŸΌ

proud bobcat
#

What’s the consensus on kimi k2.6

median smelt
#

oh it worked

#

nvm

silent tree
#

@storm dust I'm sorry.

light sleet
#

Yo night fury

storm dust
#

blocked people

light sleet
#

Why u so rare to achieve

storm dust
#

i wont unblock you again

rocky geyser
#

Whats yalls fav ai model atm?

light sleet
echo aurora
rocky geyser
median smelt
#

I didnt get image 2 yet

rocky geyser
#

lmao

median smelt
#

which is why it should be added to ARENAAAAAA!!

light sleet
median smelt
rocky geyser
light sleet
#

look how everyone's model requesting image 2 now

rocky geyser
#

shouldve used google flow there wouldve been no gemini tag πŸ˜”

light sleet
#

Everyone model request it

light sleet
light sleet
#

πŸ˜”

#

Shouldve use Google AI Studio

rocky geyser
# light sleet shouldve used arena

Nah wouldve taken too long, dont really like image generation there, im fine with google flow since it gives unlimited nano banana 2 generations

light sleet
#

Okok

rocky geyser
light sleet
#

Gemini 3.1 Pro is

rugged abyss
#

When Kimi k2.6 :)

vale quest
#

@echo aurora Can you start slowly releasing models like opus 4.5 and stuff

still musk
median smelt
#

First, image 2

vale quest
#

And announce when u do it

wicked sage
#

where the hell

#

wheres kimi k2.6

#

mad

vale quest
#

Like say

Added muse spark to direct chat and side by side

#

Inside announcement

copper ferry
#

#ukunderground #dancehallmusic #newmusic
Based Lychee coming with this Big gy4t song DA DON OF LONDON prod. Mastercard2k

Thanks for Kunal for pretending to be Bobz and Thankyou Richard for being my cameraman!

Go listen to it on Soundcloud - https://soundcloud.com/basedlychee/sets/da-don-of-london

Like and Sub! Make sure 2 comment!

β–Ά Play video
still musk
#

The direct chat model would be: gpt-image-2-high-fidelity?

echo aurora
still musk
proud bobcat
#

qwen crushing coding

formal delta
proud bobcat
#

i wonder how good kimi will be

proud bobcat
#

i have yet to test it myself

formal delta
#

Ah, I thought you already tested it that’s why I asked

hoary elbow
#

I’m testing qwen 3.6 plus right now, but I wanna know which one’s better the max version or the plus version

light sleet
#

Pineapple if u have time can u come to Chatroom

#

And set ur name to Pineapple or anything

echo aurora
urban herald
echo aurora
honest verge
#

Gpt image V2 is not out

#

I don't know what everyone is talking about

median smelt
#

Its partially out

#

OpenAI just didnt make an announcement

gleaming island
echo aurora
gleaming island
#

please

#

I'm hungry

median smelt
#

Finally bypassed nick wilde

honest verge
median smelt
median smelt
echo aurora
formal delta
#

Why do I get session not found everytime... bruh

gleaming island
wicked sage
gleaming island
#

who owns it then 😑

wicked sage
#

John L LMArena

gleaming island
#

they should add image support to claude opus 4.7 thinking

echo aurora
wicked sage
#

jokes aside its made by these people i think

#

Wei-Lin Chiang, Anastasios N. Angelopoulos, Ion Stoica

echo aurora
wicked sage
gleaming island
#

πŸ₯Ί

formal delta
echo aurora
gleaming island
#

@urban herald hi

gleaming island
urban herald
echo aurora
gleaming island
formal delta
urban herald
gleaming island
gleaming island
daring sinew
#

Does the cooldown keep increasing everytime you reach the limit?

Yesterday I reached the limit and got 15 minutes I think, I reached again today morning it gave me 25-30 mins... I reached limit once again and now it says come back in 50 minutes... Will this keep increasing over time???

How does this work?

rocky geyser
#

Never reached it 😭

daring sinew
#

You have reached your rate limit for gpt-image-1.5-high-fidelity. Please try again in 46 minutes.

Reached the limit 3 times in 2 days

rocky geyser
#

how many can you generate before the limit

#

Btw why would you chose gpt 1.5 instead of nano banana 2

daring sinew
#

Idk I made like a couple dozen? Maybe like 15-20 images...

daring sinew
rocky geyser
rocky geyser
surreal zephyr
rocky geyser
loud herald
loud herald
#

Just battle

rocky geyser
#

Oh okay

#

Well then id have to be lucky to get it

loud herald
#

This is what I mean @surreal zephyr ^^^

#

That is nano2

surreal zephyr
daring sinew
loud herald
#

And I dont even like generative stuff

hoary elbow
loud herald
#

I wish I was still friends with that dude, thbat was a great idea herrsheys

surreal zephyr
loud herald
#

For qwen being a decently sized model its good its just not quite on consumer hardware yet

loud herald
cedar citrus
#

hello guys

#

im back

#

which one is better kimi or qwen 3.6

rocky geyser
#

Kimi k 2.6

cedar citrus
#

i agree

rocky geyser
#

:D

cedar citrus
#

why did arena didnt add kimi yet

hoary elbow
#

Is 3.6+ better than max

cedar citrus
echo aurora
rigid pasture
cedar citrus
#

i hope its on side by side and direct

rocky geyser
still musk
cedar citrus
sharp iron
rocky geyser
cedar citrus
#

weird

rocky geyser
#

wdym?

cedar citrus
#

bruh

#

it was hidden

abstract plinth
#

hey is muse spark not available in side by side??

cedar citrus
#

at th ebottom

rocky geyser
#

Lmao

cedar citrus
#

i now have to wait kimi 2.6 drop on nvidia nim api

#

nvidia nim api is the goat

#

free AI api

#

just 40 rpm

#

rate limiting

loud herald
# rocky geyser

Pretty sure they added as placeholder for now and will ping when its actually on it

loud herald
#

Ah

cedar citrus
#

i tested it

loud herald
#

No clue why theres no ping then

cedar citrus
#

lol

loud herald
#

I find datasets so intriguing

surreal zephyr
loud herald
cedar citrus
loud herald
#

1M is high

cedar citrus
#

also its standart

surreal zephyr
cedar citrus
#

actually good model

surreal zephyr
cedar citrus
#

im not joking

loud herald
surreal zephyr
#

qwen has 1m even

cedar citrus
#

i just found my go to cheap opus level model

cedar citrus
#

but

loud herald
cedar citrus
#

kimi is cheaper

#

i am using it right now

#

to review my database

loud herald
#

I feel like its just gonna be another GLM 5.1

cedar citrus
#

its reasoning is going well

cedar citrus
#

i used glm 5.1 and its HORRIBLE

cedar citrus
#

i vomit when i see it

desert pendant
#

Glm 4.7 was better

loud herald
#

Its not the worst but it aint good

#

\

desert pendant
#

Used it for a few tasks but when they released Glm 5.1, they made it worse and they also made Glm 4.7 bad 😭

cedar citrus
#

weird my kilo code agent on kimi just stopped randomly

loud herald
# rocky geyser

Honestly, it shouldnt even have to think THAT much to respond like it did

surreal zephyr
# rocky geyser

saying hi to gpt 5.4 pro extended is better (80$ lost in compute πŸ₯€ )

cedar citrus
#

guys im not joking

#

its talking

loud herald
cedar citrus
#

exactly likke claude

cedar citrus
#

its horrible optimized because of glm 5.1

desert pendant
cedar citrus
#

im never touching a z ai model again

surreal zephyr
#

just use gpt

desert pendant
#

wow πŸ’”

cedar citrus
#

i wasted 2 dollars on that crap

sweet knot
#

Wait, what happened to the LLMs like ChatGPT 5.4, 5.5, Claude 4.7 Opus, and Meta’s newest proprietary model? Where did they suddenly disappear to?

Why have they been left out? Why can’t they be included in the comparison? This seems very suspicious and not quite right.

surreal zephyr
#

5.5 isnt even a real model

cedar citrus
loud herald
signal pelican
#

In case Ampro says anything, ignore him.

surreal zephyr
#

😭

surreal zephyr
cedar citrus
#

guys

cedar citrus
#

im not wasting 40 dollars

#

on claude and chatgpt

desert pendant
#

me too ✌️

sweet knot
cedar citrus
#

for me thats 200 reais

echo aurora
sharp iron
#

is there a way to export my chat

loud herald
#

I used codex 5.3 to build this agent and I just simply do the training

cedar citrus
#

but i cant chat with them

#

they are horrible at chatting

cedar citrus
#

thats why im in claude

echo aurora
cedar citrus
#

but kimi is doing great for me]

loud herald
desert pendant
desert pendant
#

which universe are you at?

signal pelican
#

sure it does

cedar citrus
desert pendant
#

gpt always fail to do good codes for me

#

✌️

cedar citrus
loud herald
cedar citrus
#

gpt can be as good

#

but claude is better

desert pendant
#

tho gpt 5.4 and 5.2 are good

cedar citrus
#

the problem with claude is the pricing

sweet knot
cedar citrus
#

if they fixed that

loud herald
cedar citrus
#

we would be rich

loud herald
#

Not frontend but general coding its the best

cedar citrus
#

ima dm you smt

loud herald
#

Alright

cedar citrus
#

i cant send this in chat

sharp iron
cedar citrus
#

pineapple will kill me

desert pendant
#

not my type

echo aurora
loud herald
loud herald
desert pendant
#

yo pineapple

#

the multi agent mode is good

echo aurora
surreal zephyr
cedar citrus
desert pendant
#

but sometimes it give error

desert pendant
#

the agent mode thing

echo aurora
last trench
#

Bro just give me gpt image 2 model, I wanna make some neural slope images πŸ’€πŸ’€πŸ’€

cedar citrus
cedar citrus
#

its too big

sweet knot
echo aurora
light sleet
#

Any chances you were involved in the drink pineapple juice crime?

echo aurora
light sleet
#

Hm

desert pendant
cedar citrus
#

fixes

#

he found

#

Here is the code review for v1 (nous/ and training/).


Critical Issues

1. Broken Causal Conv State During Autoregressive Generation

File: nous/generation.py (inference) / nous/model.py (SelectiveSSM.forward)

During generation, model() is called with a single token (next_token shape [1, 1]). In SelectiveSSM.forward, the new conv hidden state is computed as:

new_conv_hidden = x_proj.transpose(1, 2)[:, :, -(self.d_conv - 1) :]

When seq_len = 1 and d_conv = 4, slicing [-3:] on a length-1 tensor returns an empty tensor (shape [batch, d_inner, 0]). This empty tensor is passed back as working_memory for the next step. The 1D convolution's history is effectively reset every token, so the model loses local context across generation steps.

Fix: Maintain a fixed-size conv state buffer (length d_conv - 1) and shift-append new tokens, rather than slicing x_proj which only works when seq_len >= d_conv - 1.


Moderate Issues

2. deep_memory_init Buffer Is Never Used

File: nous/model.py lines 158–161

deep_memory = torch.zeros(
    config.num_layers, config.batch_size, config.hidden_size
)
self.register_buffer("deep_memory_init", deep_memory)

forward() ignores this buffer and creates fresh zeros when deep_memory is None. It also hardcodes batch_size into a buffer shape, which is unnecessary. This is dead code.

Fix: Either remove it, or use it to initialize deep_memory in forward().

3. get_dominant_position() Hardcodes Slice Indices

File: nous/alignment.py lines 96–117

if alignment[256:384].norm() > alignment[384:].norm():

This assumes alignment_size >= 448. The default is 512, so it works out of the box, but if alignment_size is ever reduced below 448, this will raise an IndexError.

Fix: Derive slice sizes dynamically from alignment_size.


#

Minor Issues
4. ignore_index=0 Assumes Token 0 Is Padding
File: training/loss.py line 29

ignore_index=0,
This depends entirely on the SentencePiece tokenizer configuration. If ID 0 is a real vocabulary token (e.g., <unk> in some SPM setups), the model will never learn to predict it.

Fix: Make pad_token_id configurable in NOUSConfig and pass it through.

  1. Extremely Slow SSM Implementation
    File: nous/model.py (_selective_scan)

The scan is a pure Python for loop over seq_len (up to 2048). On CUDA, this will be orders of magnitude slower than it should be. This is a performance bottleneck, not a correctness bug, but it makes v1 impractical to train at scale.

Fix: Use a vectorized scan kernel or consider a standard Mamba/CUDA implementation.

Design Observations (Not Bugs)
combined_loss phase weighting: Phase 2 uses 0.1 * next_token + 0.9 * coherence, which heavily suppresses language modeling. If this is intentional, fine β€” but it means the model may stop learning to predict tokens well during crystallization.
deep_memory injection: Every layer receives the same scalar-broadcasted final-layer mean (x.mean(dim=1)). If you intended layer-specific memory, this is not doing that.
alignment_update_interval = 10_000_000: This is so large that the alignment layer will likely never update during a standard training run.
Bottom line: The biggest problem is the broken conv state during generation, which will cause degraded/incoherent outputs. The dead deep_memory_init buffer and hardcoded alignment slices should also be cleaned up.

#

thats it

crystal mica
#

so gemini 3.1 pro, claude opus and etc still dont appear

opaque maple
#

does anyone want to play my game?

#

oh

#

sorry

#

wrong

shy mason
cedar citrus
#

bro got wrong server

light sleet
#

lol

#

bro got wrong server

shy mason
#

lol

#

bro got wrong server

opaque maple
#

lol

#

bro got wrong server

cedar citrus
#

bro

silent tree
#

lol

cedar citrus
#

lol got wrong server

opaque maple
#

server

cedar citrus
#

server got wrong bro

opaque maple
#

bro got wrong bro

silent tree
#

wrong got bro server

cedar citrus
#

got wrong server bro

opaque maple
#

got bro server wrong?

silent tree
#

server got bro wrong

opaque maple
#

bro got wrong lol

silent tree
#

got server wrong bro

cedar citrus
#

wrong did server get the bro out of my tongue

silent tree
#

lol got wrong bro server

#

bro lol got wrong pineapple server

opaque maple
#

lol bro got server watermelon wrong

shy mason
#

got wrong bro server

light sleet
#

bro wrong server got

echo aurora
#

....

desert pendant
#

...

#

server wrong got bro?

urban herald