logoalt Hacker News

Two stealth models just appeared on the leaderboard, both ahead of Fable 5.2

57 points • by cheptsov • today at 11:03 AM • 13 comments • view on HN

Comments

irthomasthomas • today at 2:21 PM

Deepseek V5 trained on Huawei Ascend?

barrenko • today at 2:17 PM

This Thanksgiving period will be a cambrian explosion of capable models.

alex7o • today at 2:13 PM

Glm 5.5 I bet, but still too early to tell

wangguode • today at 2:01 PM

[flagged]

skinfaxi • today at 1:58 PM

"Random AI lab makes up a new benchmark that puts at least one of its models ahead of Fable 5.2"

akutlay • today at 2:07 PM

It is interesting that the Kimi model gets a very high “slop score” but “professional critiques” gave it a pretty high score. Do critiques have no taste or the “slop” is not slop anymore?

➕ show 1 reply
_s_a_m_ • today at 1:17 PM

some obscure design benchmark

➕ show 1 reply
cronin101 • today at 2:01 PM

Saying this when none of the gallery examples show the new model's output is the ultimate "trust me, bro" move.

➕ show 2 replies
somelamer567 • today at 2:08 PM

[flagged]

➕ show 3 replies