Pancreatic cancer is the one of the deadliest of all malignancies. The five year survival rate for patients with this disease is 3-5%. Thus, there is a compelling need for novel therapeutic strategies to improve the clinical outcome for patients with pancreatic cancer. Several groups have demonstrated for other types of solid tumors that early passage human tumor xenograft models can be used to define some genetic and molecular characteristics of specific human tumors. Published studies also suggest that murine tumorgraft models (early passage xenografts derived from direct implantation of primary tumor specimens) may be useful in identifying compounds with efficacy against specific tumor types. Because pancreatic cancer is a fatal disease and few well-characterized model systems are available for translational research, we developed and characterized a panel of pancreatic tumorgraft models for biological evaluation and therapeutic drug testing. Of the 41 primary tumor specimens implanted subcutaneously into mice, 35 produced viable tumorgraft models. We document the fidelity of histological and morphological characteristics and of KRAS mutation status among primary (F0), F1, and F2 tumors for the twenty models that have progressed to the F3 generation. Importantly, our procedures produced a take rate of 85%, higher than any reported in the literature. Primary tumor specimens that failed to produce tumorgrafts were those that either contained <10% tumor cells or that were obtained from significantly smaller primary tumors. In view of the fidelity of characteristics of primary tumor specimens through at least the F2 generation in mice, we propose that these tumorgraft models represent a useful tool for identifying critical characteristics of pancreatic tumors and for evaluating potential therapies.