The article titled "ActiveGLAE: A Benchmark for Deep Active Learning with Transformers" by Lukas Rauch, Matthias Aßenmacher, Denis Huseljic, Moritz Wirth, Bernd Bischl and Bernhard Sick proposes the ACTIVEGLAE benchmark to establish a standardized evaluation protocol for transformer-based language models in Deep Active Learning (DAL). It addresses the lack of uniformity in experimental settings, facilitating comparisons and recommendations. The authors identify key challenges in DAL, present baseline results, and offer guidelines for researchers and practitioners.