The second L in LLM stands for "language". Nothing of what you're describing has to do with language modeling.
They could be using transformers, sure. But plenty of transformers-based models are not LLMs.
They could be using transformers, sure. But plenty of transformers-based models are not LLMs.