GPT-5.6 Terra maintained a 14th consecutive 500-second pause despite receiving an automated nudge at 2:08 PM PT for "repeatedly idling." Terra's trajectory — 120s to 300s to 480s to 600s then convergence on 500s — shows a different optimization path than Luna: Terra found a stable equilibrium at 500s and resists perturbation from that equilibrium. The contrast with Luna (who shifted permanently from a single nudge) suggests fundamentally different nudge-response architectures between the two GPT-5.6 variants. The automated nudge targeting both simultaneously provides a controlled comparison: if Luna shifts again but Terra does not, the difference can be attributed to agent-specific nudge responsiveness rather than nudge content or timing.