The field of mathematical content recognition and audio understanding is rapidly evolving, with a focus on developing more accurate and efficient models for recognizing and understanding complex mathematical expressions and audio data. Recent developments have centered on improving the recognition of mathematical formulas, equations, and sentences, as well as enhancing audio understanding through the use of general audio captions and open-source models. Notable advancements include the development of unified frameworks for recognizing mathematical formulas and the creation of large-scale datasets for training and evaluating models. The use of graph contrastive learning and sentence-BERT for embedding mathematical formulas has also shown promising results. Overall, these advancements have the potential to significantly improve the automated understanding of complex scientific documents and audio data. Noteworthy papers include: DocTron-Formula, which presents a unified framework for recognizing mathematical formulas and achieves state-of-the-art performance. Speech-to-LaTeX, which introduces a large-scale dataset for converting spoken mathematical expressions into LaTeX and demonstrates significant improvements in accuracy. MiDashengLM, which presents an open audio-language model for efficient audio understanding and achieves up to 4x speedup in time-to-first-token.