AnalysisAI Models
Jun 17, 12:32 AM
Weibo's VibeThinker-3B sparks benchmark debate
A team of nine researchers at Sina Weibo released a 14-page technical report on arXiv describing their 3B-parameter VibeThinker-3B model. The model's claimed performance has sparked widespread discussion about the validity of AI benchmarks.
·
Jun 17, 12:32 AM