not a bug, just how zenrows displays it, the \(16/\)57/\(165/\)456 cards say /mo but that's the annual billing rate, monthly is actually \(19/\)69/\(199/\)549 per their own docs. worth flagging since the whole table's premise is catching exactly this kind of thing. on the anti-bot bypass claim i'd walk that back to "depends which week you test". zenrows scores anywhere from 94 to 17% depending on which benchmark and target set you look at, so calling any one vendor best at bypass is kind of meaningless without pinning down the conditions. trial at your actual production concurrency too, not just a quick test, since success rates tank once you push real throughput. and the best overall regardless of scale pick falls apart on the piece's own numbers, mrscraper runs about $1 per 1k pages once you account for token use while zyte is $0,13 per 1k for plain http,so which one wins depends entirely on whether you're doing selector heavy scraping or just grabbing static pages