A $100,000 prize pool anchors the new ARC White-Box Estimation Challenge hosted by AIcrowd. The contest tasks participants with improving estimation algorithms for random MLPs. This follows recent research on wide random networks. Practitioners can now test their ability to quantify model behavior through a structured, competitive framework starting this week.