A tiny, transparent research specimen rather than a production workhorse. At 160M parameters, it handles only simple text tasks and has clear limitations in reasoning and knowledge depth. Its real value lies in being one of several seed variants released openly, letting researchers study how random initialization affects training dynamics and model behavior.