Based on weak-coupling anisotropic BCS theory, the temperature dependence of energy gap and the specific heat are evaluated for the MgB 2 superconductor, and the results are compared with experimental data. We show that the weak-coupling anisotropic BCS theory describes thermodynamic experimental data with high precision, 3-6%. DOI: 10.1103/PhysRevB.71.012514 PACS number͑s͒: 74.70.Ad, 74.20.Fg, 74.25.Bt A keen interest excited by discovery and experimental investigation of a new high T c superconductor MgB 2 is to a large extent associated with its dissimilarity to cuprate superconductors. The superconductivity of MgB 2 is definitely a three-dimensional effect, whereas in cuprates it is presumably two-dimensional ͑2D͒. Nevertheless, the superconducting gap in MgB 2 displays strong anisotropy. The most precise tunneling measurements by Gonnelli et al. 1 give the value 2.6 for the ratio of the gaps at two conductivity bands. On the other hand, the measured gaps are the same for the tunneling in the ab-plane and in the c-direction, indicating that they do not depend on direction within each piece of the Fermi surface.An important problem is how strong is the interaction in MgB 2 . First-principles calculations 2-4 indicate that electronphonon interaction is not weak and that the Eliashberg description is appropriate. However, anisotropy and interaction were shown to influence thermodynamics oppositely. For example, the anisotropy decreases the relative discontinuity of the specific heat at the transition point, 5,6 whereas the first correction due to interaction increases it.7 Besides, MgB 2 is a very hard material with a high value of Debye frequency, which usually correlates with a weak coupling. Therefore it is not a priori clear what is more substantial in the case of MgB 2 .The purpose of our work is to demonstrate that the anisotropy effects are more substantial at least for thermodynamic measurements. We show that, as a matter of fact, the weak coupling anisotropic BCS theory describes all known thermodynamic experimental data including the temperature dependence of the energy gap and specific heat with a high precision, 3-6%.The main features of the anisotropic weak coupling BCS model were elucidated in the early 1960s, 5,6,[8][9][10] the ultimate result being the factorization of the gap 5 ⌬͑T,k͒ = Q͑T͒͑k͒, ͑1͒ which was experimentally verified by Zavaritskii. 11 The function of angle ͑k͒ is the eigenfunction of the interaction operator V͑k , kЈ͒ corresponding to the maximal eigenvalue + . It satisfies the linear homogeneous integral equation: