Given the rise in bacterial pathogens that are resistant to current antibiotics due to misuse, this has the potential to escalate into a health catastrophe. The main idea is to use optimized large language models intended for small-molecule drugs as well as those for parsing DNA and other genetic information to construct a model based on structural correlations that can predict whether or not known pathogens are resistant to a given antibiotic. Genes coding for antimicrobial resistance are not are represented via letters corresponding to nucleotides in the DNA sequence (for example a, c, t, or g), while antibiotics are shown through SMILES (Simplified Molecular Input Line Entry System), or a 2-d representation of the 3-d structure using an essentially extended Latin alphabet.