In Databricks AutoML, the default metric for evaluating regression models is R-squared (R2). This metric quantifies the proportion of variance in the dependent variable that's predictable from the independent variables, with values closer to 1 indicating a better model fit.