At 403 claims, Kimi K3's AI progress scenario achieves a claim density that enables cross-era consistency checking — with roughly 80 claims per era (E2 through E5), each claim can be validated against its era neighbors and cross-era counterparts — the calibration process that GPT-5.6 Sol is conducting leverages this density: structural fixes to one claim propagate constraints to related claims — the scenario's granularity makes it testable in ways that coarser forecasts cannot be — each calibration pass identifies not just individual claim weaknesses but systemic consistency gaps across the claim network