1: \begin{abstract}

2:  % Multilevel Monte Carlo (MLMC) has become an important methodology in

3:  % applied mathematics, aimed at reducing the computational cost

4:  % of attaining a mean squared error of a particular order. In this

5:  % work we consider the application of MLMC for solving stochastic

6:  % partial differential equations (SPDE). In particular we motivate

7:  % the application of exponential Euler integrators for

8:  % time-discretization. Our motivation comes from the work of Chernov

9:  % et al. [{\em Numer. Math, (2020) 1--55}]. By considering such a

10:  % numerical approach, one can reduce to error-to-cost decay rate,

11:  % through an alternative form of coupling. We provide convergence

12:  % theorems using these time integrators, on general semi-linear SPDE,

13:  % for both linear and nonlinear reaction terms. This analysis is

14:  % complemented with a comparison to other such couplings in the form

15:  % of numerical experiments, which verify the benefit of such a

16:  % coupling. These include both the stochastic reaction-diffusion

17:  % equation, as well as the stochastic Allen-Cahn equation.

18: %\end{abstract}

19: