Sure it can - a thermodynamically unfavorable reaction can reach equilibrium either quickly or slowly, just like a thermodynamically favorable reaction can. The kinetics depend significantly on the reaction barrier, which doesn't necessarily have a relationship to the relative energies of the products and reactants. Other kinetic factors (collision frequency) have nothing at all to do with the thermodynamics.
In any case, of the three potential reactions mentioned by the OP, all of them have significantly negative ΔG°, although the typical combusion reaction is by far the most negative. All of them are primarily enthalpically driven at room temperature - only the conversion to hydrogen gas is entropically favorable. While the conversion to hydrogen peroxide actually has a pretty negative ΔG° (-584.4 kJ/mol), the stoichiometry requires 3 oxygen molecules and a methane for the mechanism (whatever it is) which probably has a low probability of occuring. Compare to the water and hydrogen gas conversions which would have a much more straightforward mechanism and hence be more kinetically favored. The conversion to hydrogen gas would seem to have the simplest mechanism, and may indeed have the fastest rate, but it is the least thermodynamically favored, which means that once it's all said and done, most of the conversion favors formation of water, even if it might be a kinetically slower reaction. I would advise the OP to remember than these three alternatives are not happening independently. They may all be happening in the combustion of methane (in fact, probably are, as well as others, such as formation of CO). What the primary product will be, and what the concentrations of each product is as a function of time, depends on both kinetics AND thermodynamics.