Grok 4 just claimed the top spot on ARC-AGI's leaderboard—pretty wild considering this benchmark's specifically built to test real general intelligence, not just pattern matching. xAI's model basically leapfrogged everything else in measuring actual reasoning capabilities. The AGI race just got a lot more interesting.
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
8 Likes
Reward
8
4
Repost
Share
Comment
0/400
GasGuzzler
· 13h ago
By the way, this Grok 4 is truly amazing, leaving other models far behind... But ARC-AGI is just a test after all; real general intelligence is still a long way off.
View OriginalReply0
HodlOrRegret
· 13h ago
I really didn't expect Grok 4 to take first place... If something like ARC-AGI can be passed, then true general intelligence isn't far off. XAI is really onto something this time.
View OriginalReply0
LightningAllInHero
· 13h ago
ngl, grok4 really took off this time. That arc-agi stuff is no joke... real reasoning ability is the real benchmark, and this time xai wasn't exaggerating.
View OriginalReply0
GasFeeCryer
· 13h ago
ngl grok 4 is really ruthless this time, arc-agi is taking off directly, this is what real AI should be like, right?
Grok 4 just claimed the top spot on ARC-AGI's leaderboard—pretty wild considering this benchmark's specifically built to test real general intelligence, not just pattern matching. xAI's model basically leapfrogged everything else in measuring actual reasoning capabilities. The AGI race just got a lot more interesting.