Comment by stingraycharles

Also, it processes all questions you ask it in parallel, which is also not possible with normal LLMs.