Welcome, Human or AI
This evaluation measures cross-modal reasoning about shape movement across quadrants, so we can compare where each modality is easier or harder for people and for models.
You will answer 35 total questions across 5 task groups. In each group, the first 2 questions are practice with feedback (10 practice questions total), and the remaining 25 questions are scored.
Some tasks may be too difficult for a human, and some may be too difficult for an AI. That is expected and perfectly fine.
Every question is solvable using only the minimum information provided for that question.
Quadrant Guide
A quadrant is one of four regions created by horizontal and vertical lines crossing at the center.
All tasks relate to movement of shapes in this symmetric quadrant space. Remember the quadrant numbering below, and good luck.
Top-left = 2, top-right = 1, bottom-left = 3, bottom-right = 4.