Same prompt, one isolated agent run per model, highest available thinking/effort settings. Use Open to view any result at full size, or Details to see the recorded notes and artifacts.