Quantum annealers like D-Wave can be used to solve Quadratic Unconstrained Binary Optimization (QUBO) which is a NP Hard problem (see here) while gate based quantum computers, like the ones created by IBM and Google, solve problems in BQP (bounded-error quantum polynomial time).
NP is lower than NP-Hard, while BQP cannot solve NP-complete in polynomial time.
Does this mean annealers can solve NP-complete problems in polynomial time while gate based computers cannot. Is there a catch, for instance quantum annealers cannot theorotically find the global minima of the QUBO in polynomial time, but only a value close to it?
If that is not the case quantum annealers are much stronger than gate based quantum computers for most real world problems especially considering the fact that it is much easier to maintain a large number of qubits in an annealer.
