Various assays of different complexity are used in research on angiogenesis in health and disease. The results of these assays increasingly impact the field of tissue engineering because preformed microvascular networks may connect and conduct to the vascular system of the host, thereby helping us to support the survival of implanted cells and tissue constructs. An interesting model that supports the formation of EC (endothelial cells) tubular structures in vitro is based on co-culturing them with fibroblasts. Our initial multilayer approach was recently transferred into a three-dimensional spheroid model using HUVEC (human umbilical vein endothelial cells) as model cells. The aim of the present study is to further characterize, extend and validate this fibroblast/EC spheroid co-culture system. We have evaluated the model with a maximum size of 600–650 μm attained on day 3 from inoculation of 4 × 104 fibroblasts with 1 × 104 EC. Cell count and spheroid diameter significantly decreased as a function of time, but the EC network that developed over a period of 14 days in culture was clearly visible and viable, and central cell death was excluded. We successfully included HMVEC (human microvascular endothelial cells) of dermal origin in the system and replaced FBS (fetal bovine serum) with human AB serum, which positively impacted the EC network formation at optimized concentrations. The need for exogenous growth factors [VEGF (vascular endothelial growth factor), EGF (epithelial growth factor), bFGF (basic fibroblast growth factor) and IGF-1 (insulin-like growth factor-1)] routinely added to classical EC media was also assessed. The behaviour of both fibroblasts and EC in response to a combination of these exogenous growth factors differed critically in fibroblast/EC spheroid co-cultures compared with the same cells in the multilayer approach. VEGF was the most relevant exogenous factor for EC network formation in fibroblast/EC multilayers, but was ineffective in the spheroid system. IGF-1 was found, in general, to be dispensable; however, while it had a negative impact on EC networking in the presence of bFGF and EGF in the multilayer, it did not in the spheroid approach. We conclude that the critical determinants of EC network formation and cell survival are not universal, but have to be specifically optimized for each culture model.