Clathrate hydrates are ice-like compounds that can form at high pressure and low temperature in presence of water and small molecules (usually light hydrocarbons, rare gases...). They are involved in many issues, mainly flow assurance (hydrate plugs in deep-sea pipelines), gas separation and storage (carbon dioxide capture in flue gases, storage under deep-sea sediments), air conditioning. They also are an important resource of methane on earth, and might be present on other planets. Whatever the subject, gas hydrate modelling is an important step to master for the design of industrial processes or the understanding of natural clathrate hydrates formation. But, these compounds are non stochiometric solids. Hence, the modelling of their composition is a difficult task. Indeed, if their compositions is not the same in the whole solid phase, it could be the consequences of thermodynamic or kinetic phenomena. Today, if thermodynamic modelling of pure clathrate hydrates is well known, it is not the case for mixed hydrates (involving several gases). The thermodynamic models are not able to predict with accuracy the composition of such gas hydrates. Moreover, if the composition is driven by kinetics (a new model have been published recently by Herri and Kwaterski, 2012), the use of experimental measurements for the study of gas hydrates thermodynamic equilibrium is a problem. Lately, we have been investigating gas hydrate formation in a pressurized batch reactor. In our apparatus, it is possible to measure/calculate equilibrium temperature, pressure and composition of each phases (gas, liquid and solid). In this work, we suggest a study of gas hydrates formation under low and high driving force. For that, two procedures have been set up: one by quick crystallization (high ΔT), and the other by slow crystallization (low ΔT). The experiments, using light hydrocarbons (CO2, CH4, C2H6 or C3H8), showed that the equilibrium pressure is affected by the rate of crystallization. The thermodynamic modelling also showed that the intensive parameters at equilibrium are better predicted when slow crystallization is used. So, it quite possible that clathrate hydrates may form at non-thermodynamic equilibrium, or that thermodynamic modelling of intensive parameters is not sufficient.