Inverse models are widely used in oceanography. However, their reliability remains an open question, as comparison of inverse model results with real values of ocean parameters is difficult due to insufficient knowledge of the latter. The feasibility of extracting the ocean general circulation, mixing rates, as well as air-sea heat and freshwater fluxes from hydrographic data is studied by applying an inverse model to the CME (Community Modeling Effort) results where both the physics and parameter values are known. The inverse model assumes approximate thermal wind balance and steady state conservation laws for mass, heat, and salt, assumptions satisfied by the GCM ocean although the residuals in the tracer conservation equations are comparable to the diffusion terms in the deep ocean. Effects of errors in these equations on inverse model solutions for different variables are studied in detail. A surface layer model is designed to estimate the air-sea heat and freshwater fluxes and the results are compared to their “true” values. Experiments on various parameterizations of different variables are carried out in the hope of getting some guidance in applying the inverse model to the real ocean. The inverse model estimates for horizontal circulation are relatively robust—they are consistent with the GCM ocean circulations in most of the experiments, and effects of equation errors are more pronounced in the estimates for diffusivity and air-sea fluxes. Residuals in the equations are noisy and resemble a random distribution. In such cases, the estimates for all the parameters are very close to their true values. The conclusions of this work are different from previous works, and the discrepancies are explained.