A high degree of heterogeneity has been observed amongst Irish beef farms, with a diverse range of production systems employing different practices and technologies. Such variation can compromise the estimates obtained when stochastic frontier analysis is used to estimate the frontier under which farms in the sector operate, since it relies on the assumption that all farms operate under the same technology. A latent class stochastic frontier model is implemented using an unbalanced panel dataset constructed from farm level data for Irish beef farms between the years 2000 and 2013, in order to identify different technologies. Results obtained suggest that a single frontier model overestimates technical inefficiency compared to the model where technology heterogeneity is taken into account. Overall results highlight the importance of correctly addressing technology heterogeneity in order to obtain reliable technical efficiency measures; and the comparison of the main characteristics for different classes identified suggest the need of targeted policy measures.