← Back to feed Fashion & Style

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

Hacker News Best 23 June 2026 20h ago
VibeThinker: 3B param model that beats Opus 4.5 on reasoning with novel SFT+GRPO
60
Relevance
3/25
Freshness
23/25
Authority
18/20
Brand Signal
14/15
Depth
2/15
Relevance Freshness Authority Brand Depth
Article URL: https://arxiv.org/abs/2606.16140 Comments URL: https://news.ycombinator.com/item?id=48639240 Points: 368 # Comments: 193
Read Full Article → Hacker News Best ↗