Eh
If I program something to always reply “2” when you ask it “how many [thing] in [thing]?” It’s not really good at counting. Could it be good? Sure. But that’s not what it was designed to do.
Similarly, LLMs were not designed to count things. So it’s unsurprising when they get such an answer wrong.
Assume it is not. If you’re asking an LLM for information you don’t understand, you’re going to have a bad time. It’s not a learning tool, and using it as such is a terrible idea.
If you want to use it for search, don’t just take it at face value. Click into its sources, and verify the information.