Designing an integrated, flexible and cost effective Fischer-Tropsch process for the conversion of CO/CO2 mixtures to hydrocarbons as chemical feedstocks or sustainable aviation fuels is of great interest to the chemical industry. We demonstrate that using Ru as catalyst CO/CO2 mixtures can be flexibly transformed into hydrocarbons suitable for further processing within a single process step. The selectivity of this process to the various hydrocarbons, expressed by the chain growth probability, depends on the used CO/CO2 mixture and the operating conditions. Explicitly modeling this dependence is a challenging task. We propose a Bayesian model based on mechanistic considerations and parameterized using Bayesian inference. Our Bayesian design of experiments reveals a strong temperature dependence as well as an inhibiting effect of CO2 on chain growth. The resulting model possesses high predictive power and allows for consistent uncertainty quantification of the chain growth probability even outside of the parameterized range. This makes it useful for the design and operation of innovative, flexible and cost effective CO/CO2 Fischer-Tropsch processes.