The modern theoretical predictions on the LambdaSSigma nature of the ground state of CuCl2 have led to different answers, depending on the type (DFT-based or ab initio) and the quality of the electronic correlation treatment; for this reason the X (2)Pi(g)-(2)Sigma(g)(+) transition energy has been predicted to range from -1856 to +5887 cm(-1). The physical problem at hand lies in the difficulty of accurately describing the orientation of the 3d hole on the central Cu2+(3d(9))/Cu+(3d(9)4s(1)) ion (in the field of both chlorine ions), which implies the need of the most sophisticated nondynamic and dynamic electronic correlation treatments. We report here ab initio benchmark calculations using especially developed basis sets to study, at the CASSCF+CASPT2 and CASSCF+ACPF levels, the transition energy as well as the corresponding equilibrium geometries. The spin-orbit (SO) effects of both atoms were included in a second step through the effective Hamiltonian formalism, using the calibrated SO effective potentials developed by the Stuttgart group. Without SO at the CASSCF+ACPF level, the ground state is X (2)Pi(g) but the vertical transition energy to the (2)Sigma(g)(+) is only 99 cm(-1) at 3.95 a.u. The inclusion of the SO effects leads to a Omega=1/2 (59% (2)Pi(g),41% (2)Sigma(g)(+)) ground state, in contradiction with the Omega experimental value of 3/2. In a last step we show that the SO effects (and therefore the final Omega ordering) are critically dependent on the LambdaSSigma electronic energies, so that it is not impossible that the Omega ordering is actually changed. For theoreticians interest in this matter is not purely academic, since many properties of organometallic complexes are linked to such delicate physical effects. (C) 2004 American Institute of Physics.