The closed models aren't giving you the real thinking traces, though.

It’s wild the people don’t want to see the thinking

Those traces are just summaries of the reasoning.