The experimental family designed to train LLMs to understand sound natively.
Note You can demo the model live here!