With the advent of the software tool AlphaFold, the field of bioinformatics and structural biology has made extraordinary progress. The task of predicting protein structure, which previously required months of intensive laboratory work, can now be accomplished in a matter of hours thanks to the performance of AlphaFold2. Within the scope of this bachelor's thesis, the student will become acquainted not only with AlphaFold2 but also with other tools such as ESMFold and OmegaFold. The study will focus on proteins that were added to the Protein Data Bank (PDB) database in the previous year and therefore could not be utilized in the training data of these tools. The primary objective of the thesis is to compare the quality of these three tools and identify their strengths and weaknesses. A secondary objective is to identify proteins where the predictions from different tools vary the most. Another goal is to find proteins with the highest sequence similarity where, nonetheless, the predictions (when using the same tool) significantly differ. In this way, the thesis will contribute to a better understanding of the performance and limitations of these tools in various contexts.