Looks like this (relatively) new system of flagging needs to be explained a bit. As you all know, things can be flagged in two different ways now. But the difference is not obvious. So let's look at them from the moderator/admin point of view to understand what they do.
For those, who witnessed the DailyReminders invasion, it's not that big of a secret that certain amount of spam flags makes things disappear. Let's look at the spam queue, where they all go, shall we?
As you can see, most of the queue is filled by your friendly neighborhood bot. You might also notice that there's a lot of it. And indeed, it fills up very quick. Every message that goes outside of bot protection algorithms goes stuffed there. That includes, for example, posting too fast and posting links as level 1. So it's usually full of spambot nonsense.
Do note, that when used on a opening post of a thread, spam flags don't make it disappear, but make the thread unbumpable instead.
General usecase for this type of flag is "This should just disappear without a trace". Which is usually applicable to spambots.
Abuse flag acts differently, and, in my opinion, was ill named. When there's enough abuse flags, moderators and admins recieve a special notification.
This should be used when you think that user's behaviour warrants a warning or even a ban. This is an analog of the old system.
Speaking of the old system… It was a mess. A whole bunch of messages that usually led to one and the same post. A royal pain to navigate and properly react on. The ability to create custom messages is mostly redundant. Most of them were "spam" anyway. And for the ones that weren't, trust me, we're capable of seeing the reason with our own eyes.
Now go ahead with your newfound knowledge and flag responsibly!