it uses the word "some". not "all".
They are simply saying that some stack may be vulnerable, some may not be. Miniupnpd is one specific stack that is not vulnerable, they specify it by name.
Why would they confirm a model vulnerable if they didn't test it?
Because a lot of these security outfits make assumptions without actually testing it, wanting to rush out their finding in the name of glory and fame. I've seen another such security analysis that was reporting issues which had actually been patched by Asus over a year ago at the time.
There is a reason why Linus Torvalds has very little respect for many of these firms, and I share his opinion: too many of these are more about marketing than actual technical analysis and fact reporting. This link is just one instance where he voiced his opinion on the matter, I've seen others in the past.
Asus did the right thing being better safe then sorry imo.
Which is the thing I'm telling you: Asus didn't make any change to the firmware code... The changelog simply did not reflect reality.