Computational thinking has become a key cross-curricular competence in 21st century education, offering cognitive tools that enable learners to analyze, decompose, and solve problems in logical, creative, and efficient ways. Its value goes beyond computer science, extending to multiple disciplines and fostering critical thinking, autonomy, and the ability to transfer knowledge across diverse contexts.

In this paper, we present COMATH, an assessment tool for 9–14-year-old students' computational thinking developed within a competence-based framework and grounded in empirical evidence. The tool has been piloted in two phases during 2023-2025 across six countries, involving over 6,300 students and 100 teachers. It relies on Bebras-style tasks and quantitative and qualitative data analysis to provide a rich, formative, and improvement-oriented evaluation.

Assessing computational thinking presents methodological challenges that require a combination of qualitative and quantitative instruments, as well as the integration of teachers’ perspectives. In this regard, COMATH not only helps to assess students’ competence levels but also offers meaningful feedback to inform more effective pedagogical interventions. Longitudinal data analysis supports the identification of learning patterns, the detection of recurring difficulties, and informed decision-making to adapt teaching strategies.

Moreover, the project promotes an inclusive and equitable vision of digital competence, aligned with the principles of formative assessment and the holistic development of students. We share preliminary findings that highlight the potential of COMATH as a tool to transform assessment into a meaningful process focused on learning and continuous improvement.