A continuous time Markov Chain (Gillespie) has no time "step". The time interval to the next event is determined by an exponential probability density whose parameter is the sum of all rule activities. Please see section 3.4 of the manual
(-in-progress...) and appendix D therein, or any number of review papers on the Gillespie algorithm. Why do you think you need to set a time step?