I'm assuming you want eigen-values. It's Hermitian and almost block-diagonal. Assuming γ1 is small, you should be able to get a good estimate, if not exact answer. A two-by-two Hermitian matrix can be diagonalized exactly by assuming that it is a rotation of a diagonal matrix by some angle theta. Solve for that angle for each of the 2x2 blocks, perform the rotations, and see what it does to your off-diagonal terms. Odds are, you'll end up with some factors proportional to γ1 scattered around. From there, your best bet is probably to go with perturbation theory to get an estimate.
Edit: Might be worth a try to use γ1 as your pivot and seeing if that leaves you with a block-diagonal matrix, but I somehow doubt it.