The aim of this study was to assess the capability of small animal PET (SA-PET) devices to image four mice simultaneously to improve the throughput of SA-PET experiments in cancer research. A customized bed was designed to image up to four mice simultaneously. This bed can easily replace the bed provided by the manufacturer and is connected to an anaesthesia device. A mouse-sized phantom was imaged, mimicking simultaneous imaging of four mice with computation of recovery coefficients and spillover ratios (SORs). In addition, eight mice bearing subcutaneous tumours (human embryonal carcinoma, n=22 tumours) were simultaneously imaged in groups of four on an Inveon SA-PET scanner after injection of F-fluoro-D-glucose. Tumour activity (Bq/ml), as determined by the SA-PET, was compared with ex-vivo counting. For a 5-mm rod, recovery coefficients were 1.15 and 1.05 for a phantom imaged at the central field of view or off-centred on the customized bed, respectively. SORair and SORwater were 0.05 and 0.04 for a phantom imaged alone and 0.15 and 0.06 for a phantom imaged with three additional scatter sources, respectively. Correlation between SA-PET and ex-vivo quantification was good (r=0.91, P<0.0001). The mean ratio of PET quantitative data and ex-vivo counting was equal to 0.9 (95% confidence interval: 0.70-1.09). New generation SA-PET may be suitable for simultaneously imaging four tumour-bearing mice, although improvement in scatter correction efficiency appears necessary. The type of customized bed developed in this study could be easily adapted to other large-bore SA-PET scanners.