Using data from the international Teacher Education and Development Study: Learning to Teach Mathematics (TEDS-M), the measurement equivalence of teachers' beliefs across countries is investigated for the case of "mathematics-as-a fixed-ability". Measurement equivalence is a crucial topic in all international large-scale assessments and widely undervalued for non-achievement measures. Although it is well known that cultural differences may affect response styles or meaning of constructs, and although comparability is often not specifically examined, international large-scale assessment results are typically compared across countries in terms of scale means. Full scalar invariance is typically unattainable in such international assessments due to their sheer size and local item dependence issues. To deal with both challenges, a Bayesian approach is introduced that uses informative priors to define elastic equivalence constraints to investigate the comparability of TEDS-M future teachers' beliefs under a working assumption of approximate measurement invariance. Substantively, a clear pattern of three groups of countries emerged that corroborates preliminary findings in other studies for other teacher beliefs. Future mathematics teachers in Thailand, Malaysia and the Philippines believed more strongly that mathematics is a fixed ability, whereas the USA, Chile, Norway, Germany and Switzerland believed less strongly in this. Taiwan, Singapore, Poland and Russia took the middle ground.