On the CCLR 40/40 rating list Hannibal does not fare too well, ranking number 13 at 3114 and on the CEGT 40/120 list rated number 27 at 2694. On the IPON list Hannibal can't be tested as it crashes too often in the Shredder Classic GUI used by the site for testing.
I decided to run my own test though it was not very scientific because I held a G4 tournament with Hannibal, Houdini 1.5, Stockfish 6, Komodo 8, Fritz 12 and Gull 3. Hannibal lost all it's games. Of course, engines are designed to score well at blitz because that's how engine rating lists are complied and they may do better at slower time controls, or so I have been told. Even so, I doubt Hannibal would stand a chance against Stockfish or Komosd at any time control.
In this game against Houdini 1.5 it took a risk at move 29, the kind a human tactician might, and the resulting position were, as Jeremy Silman would say, unbalanced...big time. Stockfish and Komodo managed to thread there way through the resulting complications without much trouble. For humans, the problems may not have been so easily solved. That said, the point that Hannibal is willing to sacrifice material for a K-side attack makes it a good engine to practice against or search openings for tricky stuff you can play OTB, so while it may not be good for correspondence players, OTB players may find it useful. Whether you consider it tricky or weak depends on what you're using it for.