I mean as opposed to the ones the search engines provide for "free".

I've lost count of the number of times they've cited a link and it said the exact opposite of what they "summarized" it as saying. At this point I wouldn't be surprised that search LLM got exactly what you see on the search results (that is, didn't actually load the web page) and then just makes a plausible guess based on that. But those "free" AIs are definitely not the state of the art of what you can get from one that is actually budgeted to think and actually fetch web resources. If you pay for an AI it does much better. I'm yet to catch one out in a blatant contradiction, unlike the free search engines which I've caught multiple times. YMMV, of course, because we all have different searches and experiences and we may have different thresholds for what constitutes an "error".