Reliability in grading refers to the consistency and stability of assessment outcomes when the same work is evaluated multiple times or by different assessors. It ensures that grading is fair and unbiased, providing a dependable measure of student performance across various contexts.