Context. GJ436b is the first extrasolar planet discovered that resembles Neptune in mass and radius. Two more are known (HAT-P-11b and Kepler-4b), and many more are expected to be found in the upcoming years. The particularly interesting property of Neptune-sized planets is that their mass M p and radius R p are close to theoretical M − R relations of water planets. Given M p , R p , and equilibrium temperature, however, various internal compositions are possible. Aims. A broad set of interior structure models is presented here that illustrates the dependence of internal composition and possible phases of water occurring in presumably water-rich planets, such as GJ 436b on the uncertainty in atmospheric temperature profile and mean density. We show how the set of solutions can be narrowed down if theoretical constraints from formation and model atmospheres are applied or potentially observational constraints for the atmospheric metallicity Z 1 and the tidal Love number k 2 . Methods. We model the interior by assuming either three layers (hydrogen-helium envelope, water layer, rock core) or two layers (H/He/H 2 O envelope, rocky core). For water, we use the equation of state H 2 O-REOS based on finite temperature -density functional theory -molecular dynamics (FT-DFT-MD) simulations. Results. Some admixture of H/He appears mandatory for explaining the measured radius. For the warmest considered models, the H/He mass fraction can reduce to 10 −3 , still extending over ∼0.7 R ⊕ . If water occurs, it will be essentially in the plasma phase or in the superionic phase, but not in an ice phase. Metal-free envelope models have 0.02 < k 2 < 0.2, and the core mass cannot be determined from a measurement of k 2 . In contrast, models with 0.3 < k 2 < 0.82 require high metallicities Z 1 < 0.89 in the outer envelope. The uncertainty in core mass decreases to 0.4M p , if k 2 ≥ 0.3, and further to 0.2M p , if k 2 ≥ 0.5, and core mass and Z 1 become sensitive functions of k 2 . Conclusions. To further narrow the set of solutions, a proper treatment of the atmosphere and the evolution is necessary. We encourage efforts to observationally determine the atmospheric metallicity and the Love number k 2 .