Abridged: We detail and benchmark two sophisticated chemical models developed by the Heidelberg and Bordeaux astrochemistry groups. The main goal of this study is to elaborate on a few well-described tests for state-of-the-art astrochemical codes covering a range of physical conditions and chemical processes, in particular those aimed at constraining current and future interferometric observations of protoplanetary disks. We consider three physical models: a cold molecular cloud core, a hot core, and an outer region of a T Tauri disk. Our chemical network (for both models) is based on the original gas-phase osu_03_2008 ratefile and includes gas-grain interactions and a set of surface reactions for the H-, O-, C-, S-, and N-bearing molecules. The benchmarking is performed with the increasing complexity of the considered processes: (1) the pure gas-phase chemistry, (2) the gas-phase chemistry with accretion and desorption, and (3) the full gas-grain model with surface reactions. Using atomic initial abundances with heavily depleted metals and hydrogen in its molecular form, the chemical evolution is modeled within 10^9 years. The time-dependent abundances calculated with the two chemical models are essentially the same for all considered physical cases and for all species, including the most complex polyatomic ions and organic molecules. This result however required a lot of efforts to make all necessary details consistent through the model runs, e.g. definition of the gas particle density, density of grain surface sites, the strength and shape of the UV radiation field, etc. The reference models and the benchmark setup, along with the two chemical codes and resulting time-dependent abundances are made publicly available in the Internet: http://www.mpia.de/homes/semenov/Chemistry_benchmark/home.htmlRead less <