Abstract The effects of particle volume fraction and matrix temper on the flow and fracture characteristics of a series of particle-reinforced metal matrix composites under tensile and compressive loadings have been examined. Under compressive loading, a steady-state regime is attained in which the composite flow stress is proportional to the matrix flow stress at the same level of strain. The strength enhancement associated with the particles increases with increasing particle content and the matrix hardening exponent. The trends are consistent with predictions of finite element calculations of unit cell models, treating the particles as either spheres or cylinders with unit aspect ratio. Under tensile loading, the particles crack at a rate dependent on the intrinsic strength characteristics of the particles as well as the flow characteristics of the matrix. Particle cracking causes local softening, which reduces the work hardening rate as compared with compression deformation. This lowers both the strength and the ductility. Experimental measurements have been combined with finite element calculations to develop a damage law, incorporating the effects of the matrix strength on the particle stress. The damage law has been used to simulate the tensile flow response of the composites, using appropriate cell models under either isostrain or isostress conditions. Though the trends obtained from the simulations are in qualitative agreement with the experimental results, they tend to underestimate the flow stress. In all cases, tensile fracture is preceded by the formation of a neck. The condition at the onset of necking is consistent with the Considère criterion. Differences in necking strains between the composites and the monolithic matrix alloy have been rationalized on the basis of the rate of damage accumulation.