Animals including humans must cope with immediate threat and make rapid decisions between action options to survive. Without much leeway for cognitive or motor errors, this poses a formidable computational problem. Utilizing fully-immersive virtual reality with 13 natural threats, we examined escape decisions in humans. We show that escape goals are dynamically updated according to environmental changes. The decision whether and when to escape depends on time-to-impact, threat properties, and stable personal characteristics. Its implementation integrates secondary goals such as behavioral affordances. Perturbance experiments show that the underlying decision algorithm exhibits planning properties and can integrate novel actions, in keeping with model-based planning. In contrast, rapid information-seeking and foraging-suppression, are initiated based on scalar action values, which adapt over time. Our results suggest that instead of being instinctive, stereotypical, or hardwired stimulus-response patterns, human escape decisions integrate multiple variables in a flexible computational architecture, whereas secondary defensive behaviors are controlled by simpler decision mechanisms. Taken together, we provide steps towards a computational model of how the human brain rapidly solves survival challenges.