Model explainability is the degree to which humans can understand why a model produced a particular output or decision. It may be achieved through interpretable model design, feature attribution, examples, documentation, or post-hoc explanation methods.