Biogeochemical cycles in the ocean are mediated by complex and diverse microbial communities. Over the past decade, marine ecosystem and biogeochemistry models have begun to address some of this diversity by resolving several groups of (mostly autotrophic) plankton, differentiated by biogeochemical function. Here, we review recent model approaches that are rooted in the notion that an even richer diversity is fundamental to the organization of marine microbial communities. These models begin to resolve, and address the significance of, diversity within functional groups. Seeded with diverse populations spanning prescribed regions of trait space, these simulations self-select community structure according to relative fitness in the virtual environment. Such models are suited to considering ecological questions, such as the regulation of patterns of biodiversity, and to simulating the response to changing environments. A key issue for all such models is the constraint of viable trait space and trade-offs. Size-structuring and mechanistic descriptions of energy and resource allocation at the individual level can rationalize these constraints.