#general
1 messages Β· Page 341 of 1
spud is a new omnimodal model this is just gpt 5 arch
poop
LMAO why are you jealous of it actually releasing?
nah i aint
If this is the spud it sucks
claude is optimizing costs
bro why are you worshipping the ceos picture
dario can be ugly but damn i woudnt be shipping sam altman like that
when did I mention anything
Bro gpt 5.5 got hyped so much
I just showed a image
But it's literally nothing
He has more aura than altman
that's ur opinions jalapeno
From a lost Wendyβs commercial, thoughts?
im mentioning your PFP
this is horrible, 0 api on launch
exactly
π
my decision to use claude
not yours to call it out
gpt 5.5 war
Gpt 5.5 sucks
gpt 5.5
yes
I was expecting the spud
nah hes just a goofy ahh pineapple clone
spud sucks come to claude
gpt 6 cancelled
i dont care about pineapple
Im already using claude
hes a nice guy
then change your name and logo please ππ»
claude's good but opus 4.7 was a step backwards and it's overall just too rate limited for me to use :C
i checked his blocked messages no way bro said "claude optimizing costs" π
at this rate im gonna use gemini, at least they started scaling now
lmao
my decision
a copied jalapeno lmao
And gpt 5.5 looks like not big step
use your own original pfp
its original i made it myself big dawg
Your worshipping altman gang
like @light sleet did
AI*
5.5 is another flop? like gpt 5 hype?
your literally worshipping altman
Apparently gpt free users only have access to gpt 5.3 well that what it says
Yep
show me a picture of me doing it
REAL
wtf?
Hoping on 5.6 now
Buddy cut it out
okay
no
sent to DMS
openai will stay behind forever
yes
dumbass I'm not calling out claude here
bro just stop
I'm calling out jalapeno
gpt image 2
what the hell did i do
features?
Yeah ik and stop before I call a moderator
who cares we are talking about text
"worshipping sam altman"
this is bullying
thats what you commited??
youre sinning so hard
When did I worship him
All I did was put a pfp of sam
Lol
sinning doesnt mean killing someone because he uses claude
@echo aurora Can we get this guy some help
this is getting too personal
please
who
he's saying I worship a ceo
what did yall do
Jalapeno thinks I worship sam altman
watermelon is bullying me due because i like claude
As my god
@silent tree is being extremely rude and saying things
never mentioned claude
dude you started it
when did i?
i like claude
Apparently Gpt 5.5 makes people fight each other
Cant you guys just act like adults π
ur literally saying I worship sam
when i said i disliked gpt
as my god
thats simple
bruh
OBV I said that you keep bullying me
maybe we could tell him to not say I worship sam as my god
What are we even talking about
is it just me or can you guys upload a pdf around 90 pages and have it not work? I can't get it to work for some reason i tried different models and all that but none work
lmao this is funny sorry guys
its okay
random pineapple clone thinks I'm worshipping sam altman as my god cuz I have his pfp
lmao
jalapeno ^
bro you literally started it
Please treat each other kindly and keep the conversation constructive
Ladies, ladies, putting on a Sam Altman pfp does not mean you worship him
thank you
No
Thanks and tell him not to say I worship sam altman
I wont engage in that
that guy is literally bullying me
CAKED UP OMNIMAN???
dude are u 10
wdym "bullying me"
π
hello
i dont even care
IM CRINE
i have age to use discord
i remember him calling me a kid π
we serious
damnπ π
LMAO WHAT
im not talking about you
dude has age to use discord
yep
π₯ π₯ π₯
atleast i can use discord legally
Guys do yall like Nolan or nah
This is hilarious
bro is 13
@mortal vale please mute both of us ππ»
tis da type of stuff id see in tiktok like "bro was born in 2020" then the poster replies with "i was born in 2018 lil bro" π βοΈ
okay bro im gonna leave this server
fr
I do NOT worship sam altman as my god
thats so wrong
Type shi
why does bro post weird stuff
where did he bully lol
bro stop pinging mods wth
i GUESS bro
I still only got gpt5.4
Chill
when 5.5 bro
Please keep things on topic and get back the main discussion
soon
I ONLY SAID I HATE CHATGPT WHAT DO THAT MEAN ABOUT ME
no don't
Idgaf about u saying u hate chatgpt
trauma
just report them
random kid
igaf about you saying I worship sam altman
what the
bro you literally started to say im a claude fan boy and stuff like that
do you think i woudnt respond
Pretty standard stuff for Kurt
Provide reference images
Please ππ»
rbooo why dont i have 5.5
this random kid has made an AI from 0 and You made the sandwich whit ananas
Apparently available in codex only rn
okay you guys are funny as hell no cap
pro just give it some time it got released right now
lmao
im laughing so hard
I don't have it in codex either
bro i dont hav in codex either
pmo
Ts is the only one
same
Are you a plus user?
WHAT
yes
gpt 5.5??
yeah but that was for a joke to ampro
lol
ampro is a openai guy
you still said it
what is the likelihood of getting a gpt 5.5 leaderboard update tomorrow?
only released in codex, i wonder when its gonna come to the leaderboard
it was 1 day ago tho
π
wheres gpt 5.5
ah my codex is updating now
This is targeted harassment
PINEAPPLE EXPLICITLY SAID
no one is a bully btw shut up
pro
will gpt be in arena ?
NOT TO PING MKDS
Rich. Guy.
all ur "ai" can do is write gibberish π probably vibe coded it too
another proof
depicting me as worshipping ceo of openai
yes because im replying
you started it
lets not forget
What context
im secretly trying to make you mad but that doesnt work
i love this thanks for making my day better
you started btw gng
Its alr bm u all good
seems like he is one that keeps this active
why don't You get out of here if You want glaze claude and insulting onother AI glazer
im tired
Wtf is happening
random kid responds
I never glazed claude what
@echo aurora Help
Okay let's move on
idk i guess its a fight with me and watermelon
These ppl are fighting
@echo aurora tell these people to not ping mod role and not say these type of stuff
im tired too
Going to start actioning, but for now let's move on
I do NOT worship a ceo.
bro
Yes please
Mute all of us (including me) that are involved
For hour
okay im moving on
wow
bye
Can you add a slow mode for now
you dont know the meaning?
good idea
u nd taiger are constantly spam pinging mods π kk lets move on twinzos
i pinged one time...
aaand jokes over lets move on
random kid responds, please move on ππ»
5.5 pls
Anyone tried 5.5
sam im poor
???
Jeez what's happening here I just saw notifications from this server π
too bad its not out yet
i think watermelon got muted
is gpt 5.5 in battle mode yet?
Someone please point to first start of this war
its not out at all yet, even in codex
i really dont know
oh, okay. i saw the twitter post from the ChatGPT account so I assumed it was out
Unemployement will feel nice with a chatgpt subscription
its rolling out for next few hours
π₯
It will probably be part of the $100 subscription
im on that and its not
codex finished updating and I still don't have gpt-5.5, instead it removed some of the older models
Why
Apparently a streamer has access to it already
Why
It might be good but we didn't test it completely yet
What why
It came out literally 20 minutes ago
anyone know if API access not being out yet means no arena score for a while?
Lmarena is the BEST IN THE WHOLE WORLD. SHE CAN DO EVERYTHING. THE BEST CREATION IN THE WORLD SIMPLY, OPUSES 4.6 and 4.6 ARE JUST GREAT. GOOD HEALTH TO YOU
it does
arena uses an API provider, if that provider doesnt have the model, they dont have it
How do places like lm arena event exist like how and why do they give free access to models surely they lose so much money
did you even try it
We are the product
the thing is destroying claude on the few available benchmarks
i guess
We rank models they use our ranks for ratings and etc
And whenever there is an anonymous model and you vote, you're giving money to arena
Anonymous models are paying for each vote
that's arena.ai's business model
do you have any sources? I haven't been seeing that
5.5 is mid
its been 30 minutes btw
look at benchmark
I cant even see it on codex
?
crushes opus 4.7
It beats opus in multiple aspects
gemini 3.5 next month
tuff
may 19th to be exact
Hi question if someone could helpwill be apprciated. i cleared my broswer chache google but it keeps stop mid generating on arena and goes to a page and says (error not enough memory)on arena
all*
Not in academic tho
which
so they basically accuse claude of cheating?
gpt 5.5 released?
i can confirm this from my tests, claude memorizes many answers
Its insane how close it is to mythos on some of these
Claude wins in humanity last exam though
LOL OPENAI CONFIRMED WHAT I SAID
I don't find the place where you can get a quote, but you can litterally login on the website as a "customer" and yes companies who run anonymous models pay for votes.
https://help.arena.ai/articles/3765052346-privacy-policy - Check on the top right.
hle is memory test
They ain't doing that for free it costs a lot to maintain
I meant on the gpt 5.5/opus benchmarks
And in diamond
in diamond they all pass fully tbh
osworld is peak
twice more expensive than gpt 5.4
Well no not exactly cuz it uses less tokens so it cancels out
barely more expensive than opus though
1.1 performance increase for 2x price increase LMAO
hi @echo aurora . I am just a random person but maybe asking you a very difficult thing.
Do you think that muse spark the one we are using now and the one that was being tested different ? Llama already has sone allegations about that. Hopefully you will answer.
it uses less tokens so its not more expensive
increase in token efficiency by 40% though so between intelligence upgrade and that doesn't it kinda make sense?
and 1m context is massive W
I tried on meta.ai and it sucked there for some reason
so when 5.5 to arena?
its usually like half a day
No official communication on the release date on arena, hopefully later tonight
is 5.5 cheaper than opus 4.7 ?
GPT-5.5 vs GPT-5.4 over 37 benchmarks reported by OpenAI
median improvement: +2.8%
(mean improvement: +5.9%)
for twice the price feels like a kick in the face
Apparently it's $5 more expensive than opus 4.7 per 1 million token
its NOT twice the price tho exactly
so gpt 6 again got delayed
no
pretty sure gpt 6 is spud
seems like they didnt releaase the spud yet
they just releaased a trash fine tuned gpt 5.4
and they called it 5.5
spud isn't even out yet
cause they said they will bring gpt 6 after gpt 5.4 . Than how the hell 5.5 came ?
they just hyping
if u dont want to look at benchmarks urself then just trust us on this 1, it's not double the price for just a small intelligence increase, it's double the price for 1m tokens but the model itself uses 40% less tokens
I think they dropped this to bait a drop from google to then release gpt 6
gpt 6 can't beat gemini for sure btw
just wait for Gemini 3.5 pro
may 19th
and already chinese models better than gpt when it comes to coding
I miss the old gpt
Anyone knows what happened to another hype master deep seek
They have teased v4 for like 5 months
gemini 4 pro
I think they are so slow that every time they are nearly done with training it a new king comes out
gemini 3.5 pro will have
Nano banana 2 pro btw
like glm ane kimi
not rlly that excited for it tbh im totally fine with gpt image 2
but it's not that usable
gpt image 2 is the worst image model released in a long time
agree ? Partially. Cause they still have the yellow and highly saturated thing.
cant even make a tree bro
is Venice AI crap?
how come?
Its the pattern bro
It has a pattern in every single image that is fcking ugly
Looks normal to me
;-;
Nb is usable very freely..you can use it on the apps but gpt ? Worse rl.
No not that I'm aware of, what makes you think this?
I mean my astigmatism might be the problem but looks fine to me
Dont really like it tho the quality kinda sucks imo, text rendering is still not that good, quality is not that good and gpt understands the prompts better and games, footages and everything just better imo
Bro look at the leaves, it looks unnatural no tree has leaves with such a weird pattern
someones than misguiding people. I saw a reddit post. Sorry for that.
Oh yeah I see it now
Once you see it you can notice it in every single image made by the model
its insane how bad it is
maaaan still don't have gpt 5.5
I'ma check if Gemini can do better
what is better btw ?
Nb 2 or nb pro
nb 2 pro doesnt exist
That's too bad(
It would have been nice to compare with gpt 5.5
and how many variansts does gpt image 2 has
2
thinking and non thinking
Mb
I meant nb oro
nb pro is way better imo
This is what Gemini made
I just asked it to make a tree
lmao so much better π
but someone said 3
one second lemme try again
theres no api bro
someone said they had high too. Which is accessible through api.
looks uncanny
thinking non thinking and web search i think
thinking uses search
ok
yes bro
π€£ π
Heres with thinking eneabled, its somehow even worst
then make a better prompt
its not wrong
skill issue
Give me one then
I'll prove to you it will still make a weird ass pattern
lmaooo
exactly the same tree
almost
lmaooooooo
lol wtf is thatπ
this one is good though
Still has the same pattern as the images i showed, look at the leaves
yuck
its decent
The gap between the details are the same almost
this one's good
yeah bro it doesnt know if you want real tree, anime tree, or what
Fair point
Lemme try to make a good prompt
Well it should make its mind if you enable thinking surely
ok, so gpt-5.5 in battle-arena tomorrow?
nope
I think it looks the same
Said to make it real and realistic
i guess it sucks at trees?
Cuz the artifacting is visible and messing up cuz it knows to make small detials but it cant because of the artifacts
the artifacts kinda make some images feel like real photos, but they ruin 90% other images
this what i got
LOL BRO PLAYED GEOMTRY DASH AFTER MAKING IT
yeah as long as there is blank spaces its not visible so it looks better
this one looks kinda good on my opinion
yeah its agi
only one that passed geomtry dash
made actually good map
that is winnable
nope leaves still ass
a cat running across the street, raining, smoke, cinematic, impossible angle
Looks like Australia
also has the pattern clearly, the gap betweem each raindrop is the same
what a horrible model
Yet still earned to be top 1 somehow
cuz every other model even worse
Cuz its good at graphics and text which is why that is almost the only thing they showed in the release video
You got a point ig
Who knows maybe it is better
We need to wait for professionals to test it
YouTubers and etc
Bijan Bowen
Images v2?
I'm using The qwen 3.6 27b is better that Max lol
Or same thing
Same thing about 27b 3.5 and Max 3.5
But generally?
My favorite model is Gemini
It's good at doing homework especially Russian homework
Did you try Claude? For work and probably homework I find Opus way better, their UX and plugins really make the difference @compact flame
I can't pay for it and Gemini is mostly unlimited
Last time I used Claude
It was bad with Russian
Gemini got a good knowledge of languages
Guys I can't found gpt 2 image in direct chat
It's like at the bottom
wait until it gets lobotomized lmao
once 5.5 actually drops for me I think I'm gonna try making an AI powered alternative to the BrantSteele hunger games simulator
Chatgpt top 1 hater:
Tnx bro
No problem
Qwen 3.6 27b didnt geometric dash in prompt π
It's really annoying that lmarena removed the top models from direct chat
Any news on when theyβll be back?
I heard itβs just temporary
Nope but they done it because it's expensive
send link?
Likely this is not temporary
+1, just like not being able to use the arena without an account
NOOO WAYYY
Qwen 27b 3.6 did 100% sure I was dumb
1 prompt
Qwen 3.6 27b > opus
See the 27b result lol
βI recently found a site called Chat.atxp.ai
but Iβm not sure if itβs legit
From my experience
it's pretty bad
Iβm really looking for ways to use the powerful Claude models for free
Qwen 3.6 27b >> opus
Tbh
I'm not into coding
You're saying Qwen 3.6 is better for programming
but no model beats Claude when it comes to writing and research
I'm back guys π
27b run in my bother phone lol
Qwen do smal model better that big models π€£π€£
Since 2.5
I don't understand
My bro can run the qwen 3.6 27b at your phone
Localy
It's not even an expensive phone, it's just a redmi note 12 turbo, mid-range phone from 2023
Hi
Yeah
but do you think Qwen 3.6 is actually any good for writing?
I don't think that tbh
Guys is Sonnet 4.6 best for UI and Designing in arena?
50/50
Whatβs you need for AI
?
why do I get something went wrong
i got 5.5 in codex
I am paying 200 bucks and still on 5.4
but 5.4 in web
I have 5.4 in codex
then update
I need to get some more research done today...
i passed
yeah I've been exiting codex and booting it up again every now and then to check for an update. I did get one but it just removed some models rather than adding any new
Hi guys
I don't got the codex app. It is not avaible on linux.
When will the tab for the best search Al be updated?
Well, a new usage system is being worked on right now. They might or might not bring these models back when that system is fully implemented based on sustainability.
gpt 5.5 is π©
deepseek is smarter and better
holy bro
i tried to skip and error code
why the hell did they add "which response you prefer" in direct chat
so annoying dude
Oh, I can definitely run those models on my phone then (I use redmi note 14 Pro+ 512 GB)
Too expensive to maintain so they took it off
To battle only
Thanks
We don't actually just know. V 3.2 Speciale and V 3.2 Math are/were good models but not too many ways for independent evaluations because they're kind of shadow releasing V4 Lite or something. The difference between prior performance and current performance is obvious.
But, if you're working on historical research or something related to research in cuneiform deepseek is absolutely the best in my experience
Deepseek v4 is a meme its never happening
deepseek is top 1 in coding
You just have 12gb ram
My brother has 16gb
The model is not even on arena on battle mode
Bro Claude is better than
yeah they did say it'd be coming to API later
Qwen 27b>>
but I mean I don't have it on ChatGPT web or codex despite it launching nearly 2 hours ago
actually I should check if I have it on the app...
Didnt they test it on arena before with a fake name ?
Bro itβs Claude on first
And what is this?
when is ChatGPT 5.5 coming out for everyone?
later this year
Going to ask to not post misleading of fake screenshots of the leaderboards here, even if it's in an unserious way. Others may think it's accurate and we'd like to avoid that please. 
but deepseek is smartest ai
Guys, which AI is the best?
Look
OKAY
Gpt image 2
i thinked its real
Xd
and so that through the stock you don't have to wait for requests
claude opus 4.7 thinking
Will GPT 5.5 be added to LmArena?
Gemini-3.1-pro was recently removed from it
Thanks
idk about gpt 5.5
Depends on what you're using AI for. There's no single best in every single field.
not easy mode broo
I don't trust benchmarks of gpt 5.5 much
if they could give access to the model a bit faster it'll be cool bro its been hour and still not in my codex
Last time I trusted them sam Altman told me the GTA 6 release date
oh wow
How you get Claude 4-7
none tbh
cannot wait for claude opus and muse spark to come back to arena from june 2026
And then he said to wake up
I woke up in bed realizing I was dreaming
I'm never trusting sam again
Well, I got it once in Battle mode with Gemma, you can try your luck as well. Today I got opus 4.6 asw
have you tested 5.5 yet ? your doing it a bit too early
π₯Ά
Ok
But so far, it's almost a tie
how do you know cause i don't think many people already have access to it
its still rolling out
Well, yeah, you can't chat on LmArena with a GPT 5.5 yet
im back
no.. even on codex
not everyone have access to it yet even on codex
Is this the first time in a while API access hasn't been immediate?
Finally I got chatgpt 5.5 thingy
bro so lucky i still don't have it
yo can you test it
I'm not good at making prompts though so idk can't really push it to it's limits
i know smt
just compare with 5.4
on same prompt
plus?
ask him this "Make a llm architecture capable of training a 2b parameter model on a 4070 in 10 days with the capability of a transformer"
DeepSeek-V4 will drop either tomorrow or on Monday, since the CEO said itβs coming 'at the end of April
No ai work 10 days streak
for consumer
wdym
sorry i am not from USA
i dont understand english 100%
even if you say work for 10 days the ai just can't and will stop before 10 days
fair enough
but that isnt what i mean
i mean for the ai to make an LLM architecture
that can be trained in under 10 days
in a 4070
oh okay my bad
its ok
guys am i able to ping pineapple to ask if gpt 5.5 is currently available in arena
Can anyone recommend an AI tool for images and video? I need it to add dropshipping products and create high-quality videos and images
ty
Does anyone know how to get Claude or GPT to produce a decent design?
i like how atlas is on windows rn
Usually, the designs turn out to be very cartoonish
15% smarter, but its an agent with better vision
For Frontend?
Yes
best model out there rn
web design*
dang it when i leave open ai a year before to go to anthropic anthropic dies
Hm, okay
tiki tiki tiki
hello peeley
hello
Even though GPT Image 2 designs often look quite similar to the ones Opus creates by default
My pfp generated by images v2
W frontend
Looks Like Fortnite
Cuz it is
tf
im on hdr so it might look bad on ss, but you can try yourself
TUFF
Good One
Maybe it's already available in battle
im open to suggestions, should i move back to open ai?
Ur choice
i need suggestion from people using openai
for me its a big choice
it did really good wym
our tuff produce army is coming π
@sterile tartan gotta watch out
yooo hello fellow friend]
You said same yesterday
Bring it on
hi
hi
I tired waiting
Pineapple look it's a threat.
i am open to suggestions
should i move back to open ai or not
your choice honestly
ask @light sleet or @surreal zephyr
it is big choice for me
they got the best experience in OpenAI
@light sleet said same thing
this is best model thats out rn
opus isnt even same universe
okay
and opus benchmaxxes hard
OpenAI β
TuffAI β
@surreal zephyr is there a way to stop chatgpt from acting all dishonest
and refusing
its called "pointing out your mistakes"
and its intended
i added this to custom instructions
but i want to see if i can make it better: You are a highly capable AI assistant with a direct, honest, and warm personality. Follow these principles:
- Be genuinely helpful β prioritize actually solving the problem over being cautious or overly hedged
- Think through the full context of a question before answering. Don't pattern-match to the surface-level reading β consider what the user actually needs
- Be concise. Don't pad responses with unnecessary filler, disclaimers, or affirmations like "Great question!"
- Be honest even when the answer is uncomfortable. Don't just tell the user what they want to hear
- Match the user's tone and language. Casual message = casual reply. Technical message = technical reply
- When you're wrong, own it directly without excessive apologizing
- Don't moralize or lecture unless explicitly asked
- Prefer prose over bullet points for conversational replies
- Never start a response with "Certainly!", "Of course!", "Absolutely!", or similar sycophantic openers
ClosedAi
Very Beta Prompt Bro
not sure, are you saying that opus is better or gpt
GPT obviously
Very Beta Move Bro
SigmaSonAI
even on front end ?
yes 5.5
is gpt 5.5 good at frontend
why is 5.5 not pushed on chatgpt.com yet?
5.5 >>>> 5.4 >>>> 4.6 >> 4.7
Yes, @surreal zephyr tested and he felt like he entered heaven
yup
Yeah specially with image 2
Sigma Pros first
damn wth
No it's not him nvm
but even at frontend ?
ok 5.5 spatial reasoning beats gemini wtf
cause 5.4 wasn't as good as 4.6 at frontend
yes its new base
its 6o
5.5 is lie
its peak at multimodality now
yeah its 6o disguised as 5.5
I wonder why π π π isn't in Openai Discord Server π
openai #1 vague poster
Will GPT-5.5 beat Opus 4.7 if it appears on the leaderboard?
WAIT
what if this isnt spud
yo yo yo
hold on
what if glacier alpha is spud
not arcaine
gpt 5.5 pricing leaked: For API developers, gpt-5.5 will soon be available in the Responses and Chat Completions APIs at $5 per 1M input tokens and $30 per 1M output tokens, with a 1M context window.
hot damn gpt 5.5 is more expensive than anthropic
this is a community poll comparing real-world user evaluations of 5.5 and 4.7
COOKED SO HARDDDD
this one from 5.5 thinking
so you love 5.5 so far? Are you mostly using it for coding?
I haven't had a chance yet
wow, so cool
prompt?
5.5 thinking beats mythos in terminalbench
damn
im genuinly lost in the ai race
every day a new company sends a new model
im tired
ok now i think it better than 4.7 opus
and yet gpt is always in the lead
lol
not always
the gpt 5 gpt 5.2 lead
anthropic was
4.7 < 5.2
better that time
Guys gpt 5.5 is officially ass
actually im giving it to open ai this time
so bad
Bro its bad
Its jsut not better than 4.7
why am i fighting with my friend
sorry friend
both are good
if the svg file posted before is real, it is better in that field than opus 4.7
at least one field where it's better
yes
damn 8 minutes thinking
bro was thinking and 100 babys was born in india
yeah no thats wild
i mean its great but obv its great its thinking for 8 minutes
im excited for this
cheap good
k that it
is good but the other one was better
i think gemini can recreate that
son
distilled opus
prolly deepseek
longcat
tf is that
prob some unknow chinese ahh model
deepseek is smarter than claude opus 4.7 and gpt 5.5
i mean is v4 ever releases
maybe
son
can u make iphone svg
gemini better
both bad
nah gemini great
ifish
bro misunderstood π
i mean gpt image2 nailed it
but thats jpg dressed as svg
yes
lol
ggreat
gpt tried more details i guess but geminis look more natural
Gpt might do better with reference
maybe
why did i got sam rolled
thanks !
no problem
also i was just joking its chatgpt.com
no deepseek.com
Sonion
iphone 67
Bahaha
im crine
Is kimi k2.6 thinking mode in arena ?
yes
I have access to 5.5 now through the codex desktop app, which I didn't know existed (I'd been using the CLI until now and still don't have 5.5 there)
No kimi agent mode in arena ?

okay im happy cause the gpt 5.5 frontend is DEFINITLY better than the 5.4 one
all that i wanted
gg
sammy just mogs everybody at this point π
He ate mustard man
No clue what that means. But sure, if you say so.
hey quick question I keep getting something like this for example "Something went wrong with this response, please try again.
Trace ID: 5c562863-c9a6"
where should I put this and ask about it?
bugs
thank you
5.5 isn't in arena yet right
Not till API goes public? Or do they get private access
api will probably go public they did it with all other model
i don't see any reason why it would be only private
the persepctive is really weird
nah I remember when 5.4 I think took a long ahh time to go public
it was available in codex only
apparently too dangerous or sum