#bot-feedback

1 messages · Page 3 of 1

crude terrace
#

Did I miss something? When I try to create videos using image-to-video it can't create the evaluation session

cedar silo
#

There isn't a way to add a video file as a reference, it's just going to be 1 image-to-video.

#

Yup, this is an experiment, meaning it's going to be random if you get the feature or not.

cedar silo
cedar silo
# abstract path What is this new function?

That's Auto-Modality, basically it'll automatically detect what modality you should be in based off of the prompt. It's an experiment so is also going to be random if you see it or not.

abstract path
#

Wow

cedar silo
#

Currently, if you're in say the Text modality but prompt "generate a picture of a dog", that's going to still keep you in Text even though based on the prompt it's pretty clear you should be in Image modality. Auto-modality will do that automatically.

cedar silo
abstract path
#

I think it was something for programming and etc
But it's more powerful Of what I thought

cedar silo
crude terrace
covert juniper
#

Lm arena best 💪

hallow kayak
#

Best

clever sapphire
#

hi, is it possible to have more than 5 attempt a day ?

remote marsh
#

Pls why is showing these

#

Is weird sometimes

#

@cedar silo

cerulean acorn
#

@chilly solar how to use video generator on LMArena.ai?

#

not on discord server?

jade ether
#

I am trying to generate a video here but cant get my way

lost steeple
#

evolution session error whyyyyyyyy😭

cedar silo
cedar silo
cedar silo
sonic pebble
#

May I have creator access so that I can train an avatar for a project?

ember dune
#

Can I generate a test to Video?

tepid gulch
#

can i generate video?

cedar silo
slim coral
sonic pebble
#

Hi LM Arena team, I’m working on a controlled text‑to‑video pipeline and need access to model‑specific generation (particularly Veo‑class models such as Veo 3.1). The public Video Arena channels are useful for comparative evaluation, but the randomized model assignment prevents deterministic generation, which is required for my workflow.
My use case involves generating identity‑stable talking‑head footage for downstream model training (HeyGen). This requires:
• consistent facial identity across multiple generations
• stable lighting and camera parameters
• repeatable outputs from the same model
• the ability to run controlled prompt experiments
The blind‑arena setup makes it impossible to isolate model behaviour or reproduce outputs, so I’m requesting access to the creator/researcher tier or whichever role enables model‑specific channels (e.g., Veo 3.1).
If this is feasible, I’d appreciate the upgrade. I’m happy to provide additional technical details if needed.
Thanks.

cedar silo
jaunty patio
#

where do i go to submit my contest content to the code arena?

sand dawn
#

Cant create video please fix the problem

worthy shale
agile ember
#

Looks like there is an issue with the video generation. All requests are stuck.

cedar silo
fiery sigil
#

My last 2 video genertion requests are still pending for over 30 min and previous one for over an hour. There seems to be an issue.

cedar silo
cold atlas
#

My video generation request started yesterday at 21:48, and it’s now 17:20—it’s still processing.

cerulean acorn
#

is the video generator official in lmarena.ai àn always appear in the website? @cedar silo

abstract path
#

My video generation doesn't work anymore, I don't know why.
I only generated 3 times yesterday.

#

It's not the photo since no photo works none works

tribal pilot
#

lm is down

abstract path
abstract path
#

Omg LM is broken

tribal pilot
#

LM IS DOWNNN

cedar silo
#

Hey @tribal pilot @abstract path would recommend you check out this Help Center for advice on what to do if you're running into this error - https://help.lmarena.ai/articles/1645798556-lmarena-how-to-something-went-wrong-with-this-response-error-message

Note that the 2nd half of the article explains how to best get us more inforamtion so we can get a better idea of what is happening. Also would ask that if you run into this problem to post in #1417174113092374689 . This thread is for feedback on the Discord bot.

cedar silo
tribal pilot
cedar silo
tribal pilot
#

WHAT THE HECK MY CHATS ARE RESETTED

#

(sorry for bad english)

cedar silo
tribal pilot
#

no it isn't a bug i just cleared a cache ...

cedar silo
abstract path
cedar silo
night laurel
#

I need Frame1 to frame2 option. When it will work on Arena?

glad stratus
#

why do i allways get this message??

cedar silo
cedar silo
abstract path
cerulean acorn
#

@chilly solar Where is the model gemini-3-pro-image-preview-2k(nano-banana-pro) in direct chat or side by side?

abstract path
#

The sad truth is that I don't think the server can stand the same here on Discord, it keeps giving an error, but it was working just now.

#

The site probably crashes due to the amount of people trying...

white swallow
tribal pilot
#

YAYYY CLAUDE OPUS 4.5 GOT FIXED

cerulean acorn
white swallow
#

oh i found it on the beta website

cerulean acorn
#

does the video generator on lmarena.ai website has a limit?

cerulean acorn
#

@white swallow