News for hackers. Stuff that matters.

A comment on VibeThinker: 3B param model that beats Opus 4.5 on reasoning with novel SFT+GRPO. Parent comment

Re:VibeThinker: 3B param model that beats Opus 4.5 on reason... by dominotw on Tuesday June 23, 2026 @12:45PM (#48647694)
> Am I right in thinking this is a tiny model which has been trained well to reason, and that's it?

i remember karpathy mentioning in dwarkesh podcast. But is reasoning really possible without all the knowledge.

[ Parent | on HN ]
Re:VibeThinker: 3B param model that beats Opus 4.5 on reason... by supern0va on Tuesday June 23, 2026 @12:58PM (#48647874)
Even Karpathy acknowledged that this would require some baseline of human knowledge. The idea wasn't pure logic/reasoning, but some subset to bootstrap from.
[ Parent | on HN ]