A compact visual reasoner that thinks before it speaks — this model uses an internal chain-of-thought process to work through image-based problems before producing an answer. At 2B parameters, it punches thoughtfully for its size but naturally has limits on complex multi-step reasoning compared to larger siblings. It handles image-text tasks with deliberate, step-by-step logic rather than reflexive responses.