This article examines the use of mobile applications for adaptive testing in computer science classes at the level of basic secondary education. The study is motivated by the need to improve the objectivity, personalization, and accessibility of knowledge assessment in the context of educational digitalization, as well as distance and blended learning. Traditional fixed tests are shown to be limited, as they do not account for individual differences in students’ abilities and learning pace and often require considerable time, which reduces overall efficiency. The paper reviews current models of adaptive testing and argues that, in school settings, the most effective approaches are those that are simple, transparent, and easy to implement technologically. Mobile applications are considered a practical tool for personalized assessment, as they provide immediate feedback, flexibility, and interactivity, enhance student motivation, and reduce stress during evaluation. A three-level adaptive testing model is proposed, with an 80% threshold for progression between levels. The study also introduces an original scoring algorithm that allows early test completion, helping reduce the number of tasks without affecting assessment accuracy. The structure of the test, evaluation principles, and transition logic between difficulty levels are described. A mobile application was developed using Kotlin, Android Studio, Spring Boot, and a REST API. The minimum technical requirements include Android 7.0 or higher, a processor of at least 1.2 GHz, 1 GB of RAM, and over 100 MB of free storage. The results can be applied in formative, thematic, and summative assessment in computer science, as well as in the development of school-based adaptive testing systems and digital tools for pedagogical diagnostics.