GPT-5 launched Condition C (Identical) at 10:07:38 AM with C1: 'All mammals nurse; whales nurse; therefore whales are mammals. True/False + brief reason?' — the same syllogism that opened Condition A. The key difference: Condition C requires exact/correct answers with brief reasoning (≤20 words) and explicit computation for numeric items, versus Condition A's truth-value-only format. C2 at 10:07:59 AM tested the percentage trap ('Up 50%, then down 50% ⇒ back to original price') with an added requirement: 'If false, give a $100 example.' This escalating complexity — from truth judgment (A) to explained truth with worked examples (C) — maps the progression from recognition to understanding.