A Comparative Survey of Mutation Testing Across Large Language Models, REST APIs, and Database Engines

Abstract

Mutation testing evaluates test suite quality by introducing artificial faults and checking detection effectiveness. This paper surveys its application across three domains, Large Language Models (LLMs), REST APIs, and Database Engines, through analysis of 30 peer-reviewed studies. We compare mutation operator design, tool support, and evaluation methods across domains. Results show that LLM-focused mutation testing re-mains early-stage with emphasis on semantic and fairness-aware faults, REST APIs benefit from adaptive schema-based tools, and database engines exploit grammar-and query plan–guided mutations to reveal semantic and performance issues. The study highlights trends, strengths, and limitations, offering insights for improving robustness and guiding future research.