06:04
2026-06-23
devclubhouse.com
large-language-models
When a 3B Model Out-Reasons Opus 4.5, Read the Fine Print
Weibo's AI group released VibeThinker-3B, a 3-billion-parameter model that achieves 94.3 on AIME and outperforms Claude Opus 4.5 on competition reasoning, but collapses on general knowledge tasks. The…