The rapid expansion of Generative Artificial Intelligence (GenAI) and advanced digital technologies has intensified demand for teacher professional development (TPD) while increasing pressure for efficient use of resources. However, most TPD evaluations focus on pedagogical outcomes rarely taking cost-efficiency, contextual factors and scalability into consideration. Furthermore, there is a lack of frameworks that connect pedagogical impact with cost structures and sustainability. This study introduces and empirically applies a Cost-Benefit Framework (CBF) designed to support systematic evaluation of TPD in Pedagogical Digital Competence (PDC). Using pre-post outcome measures and structured cost analysis within multiple-case study design in three countries, the framework enables multi-level evaluation.

At the case level, findings reveal differentiated profiles. A school-embedded model in case A, demonstrated high contextual alignment, cost-efficient delivery and positive knowledge gains, but showed limited classroom adoption. The fully online collaborative design model in case C improved engagement and intentions to use AI tools while reducing facilitation costs. Enhanced training approach with situated learning in case B resulted in stronger digital competence and self-efficacy while generating higher implementation and cognitive demands.

Across cases, distinct trade-offs and patterns emerged, particularly depth-complexity trade-offs, design-cost interactions, adoption gaps and tensions between scalability, alignment with the context and implementation practice. Overall, the CBF demonstrated its value as a decision-support tool enabling systematic and transparent evaluation of different training programmes linking outcomes with the resource use and contextual factors in the evolving landscape of digital and AI-supported education.