Computational Thinking (CT) has emerged as an essential competency in twenty-first-century education and is increasingly integrated into mathematics curricula and assessment systems globally. Despite its growing significance, limited studies have examined how CT dimensions are represented in mathematics assessments across countries and how these dimensions align with the characteristics of the Programme for International Student Assessment (PISA). This study employed a qualitative approach using comparative content analysis. The data consisted of official junior secondary mathematics assessment documents administered in 2025 in Indonesia, Malaysia, and Singapore, which were selected through purposive sampling and analyzed using a coding framework encompassing decomposition, pattern recognition, abstraction, algorithmic thinking, and evaluation. The findings reveal substantial differences in the distribution of CT dimensions across countries. Indonesia predominantly emphasizes Decomposition (33.3%), reflecting a stronger focus on contextual understanding and information extraction. Malaysia places greater emphasis on Algorithmic Thinking (26.4%) and Abstraction (26.2%), indicating a focus on procedural fluency, mathematical modeling, and systematic problem-solving. Singapore demonstrates the highest proportion of Evaluation (15.4%) alongside strong representation of Abstraction (25.6%), highlighting formal reasoning, validation, and justification processes. Comparison with PISA tasks indicates that the assessments of the three countries share similarities in contextual problem-solving, data interpretation, and mathematical reasoning; however, they provide limited opportunities for complex system modeling, causal reasoning, simulation, and evidence-based argumentation.