Extreme catastrophic risks—those with very low probability yet the potential to cause human extinction—defy conventional risk-management methods that depend on historical data. To address this gap, we introduce the Extreme Catastrophic Event Model, a structured quantitative scorecard that evaluates threats along two orthogonal dimensions: understanding (how well we grasp an event’s mechanisms, evidence base, and potential impacts) and controllability (our capacity to prevent, influence, or respond). Each dimension comprises seven factors—ranging from historical occurrence and analogous events to monitoring systems and feedback mechanisms—scored on a 0–2 scale. We applied the model to nine prototypical scenarios, including supervolcano eruptions, engineered pandemics, nuclear war, AI takeover, asteroid impacts, runaway climate change, gamma-ray bursts, alien contact, and self-replicating nanotech (“grey goo”). Cluster analysis of the resulting scores revealed two groups: one of moderately understood, partially controllable risks, and another of highly speculative, virtually uncontrollable threats. By quantifying epistemic uncertainty and intervention potential, this framework offers policymakers and researchers a tool for prioritizing resources, guiding mitigation efforts, and identifying critical research needs for existential-scale dangers.