Загрузка видео...

Не удалось загрузить видео

На главную

Gemini 4 Argon is good in webdev coding atleast thanks Synopsis for running prompt on 4 argon for me

15,569 просмотров • 6 дней назад •via X (Twitter)

Комментарии: 12

Фото профиля LiteSoul
LiteSoul6 дней назад

@abhinavflac You're not clear, is that an animation made by Argon?

Фото профиля Chetaslua
Chetaslua6 дней назад

@abhinavflac yes

Фото профиля Ziwen
Ziwen6 дней назад

@abhinavflac Feel like it’s comparable with opus 5.5

Фото профиля VT
VT6 дней назад

@abhinavflac Webdev is where most new models look good first, since a slick one-shot page is easy to show off. The real test for Argon is the second and third edit on the same codebase. Has anyone pushed it through a multi-file refactor yet?

Фото профиля Chahat Sharma
Chahat Sharma6 дней назад

@abhinavflac good compared to what though? did the same prompt go through Opus 5.5 side by side, or is it that one Argon output looked nice

Фото профиля Synopsis
Synopsis6 дней назад

No worries bro👋

Фото профиля Chetaslua
Chetaslua6 дней назад

thanks man

Фото профиля CooperCode
CooperCode5 дней назад

@abhinavflac The webdev test is a useful eval precisely because it's end-to-end: the output has to run and look right, not just pass hidden unit tests. If Argon is strong here, the interesting follow-up is whether it's better code or better taste — those are different capabilities.

Фото профиля DrstaOne
DrstaOne6 дней назад

@abhinavflac Gemini is trash

Фото профиля Chetaslua
Chetaslua6 дней назад

@abhinavflac they will have good comeback atleast for middle sized model

Фото профиля DavidSZD
DavidSZD6 дней назад

@DrstaOne @abhinavflac Opus 5.5 High has the sale cost per task on AA, and opus mog argon based on my tests.

Фото профиля DrstaOne
DrstaOne6 дней назад

@chetaslua @abhinavflac lmao

Похожие видео