Are there any distributions such that parameters have multimodal posteriors in Bayesian MCMC?

Question

I want to test a new idea I have for Bayesian MCMC, but it'd only be useful with multimodal posteriors. Are there any examples of such distributions?

For instance, with a gaussian mixture model, of course the distribution itself is multimodal. But if I were to use Bayesian metropolis hastings to find the distribution of the parameters themselves given some data, then each parameter's distribution will be unimodal.

Thanks in advance!