Aim: To parallelize scanning and save time and cut costs of preclinical studies we have designed a new hotel holding 4 rats in the HRRT, which has a spatial resolution close to that of preclinical PET scanners. In this work we test the quantitative accuracy on phantoms in the hotel using different attenuation corrections methods on the HRRT. Material and Methods: The rat hotel has 4 compartments made of acrylic plastic with an 8 mm base plate and a 3 mm half-cylinder lid. Four 50 ml syringes filled with [¹⁸F]-FDG in water were used as phantoms and scanned in the rat hotel for 20 min. on the HRRT and a high statistics speed 10 transmission scan was acquired. Three n-map processing/reconstruction methods -MAP-TR with either human head (HH) or water phantom (WP) prior and TXTV - were used and μ-maps and PET images reconstructed with each of the 3 μ-maps evaluated. Results: The μ-maps all underestimated the LAC of the acrylic plastic material as compared to CT, and the base plate thickness was underestimated. Activity concentrations were thus also underestimated: -4.6% using HH, -8.7% using TXTV and -13.8% with WP. No noteworthy local variations were found. Conclusion: We found a global underestimation of PET activity, which was within a ±5% acceptance range using MAP-TR with the human head prior and a long transmission scan (speed 10). Fine tuning HH or TXTV parameters might give further improvements.