ABSTRACT Purpose: To investigate the new Hoffer QST (Savini/Taroni) formula (HQST) and compare it to the original Hoffer Q (HQ) and 4 latest generation formulas. Setting: I.R.C.C.S. - G.B. Bietti Foundation, Rome Design: Retrospective case series Methods: Refractive outcomes of the HQST, Barrett Universal II (BUII), EVO 2.0, HQ, Kane and RBF 3.0 formulas were compared. Subgroup analysis was performed in short (<22 mm) and long (>25 mm) axial length (AL) eyes. The standard deviation (SD) of the prediction error (PE) was investigated using the heteroscedastic method. Results: A Caucasian group of eyes (n = 696), implanted with the AcrySof SN60AT (Alcon Labs), and an Asian group (n =563), implanted with the AcrySof SN60WF (Alcon Labs), were investigated. In the Caucasian group, the SD of the EVO (0.464 D), HQST (0.471 D) and Kane (0.474 D) formulas was significantly lower compared to that of the BUII (0.538 D), HQ (0.535 D) and RBF (0.503 D). In the Asian group, the heteroscedastic method did not disclose any significant difference among the SD of the 4 modern formulas (range from 0.333 to 0.346 D), whereas the SD of the HQ formula (0.384 D) was significantly higher. Compared to the original HQ, in both Caucasian and Asian groups, the HQST avoided the mean myopic PE in short eyes and the mean hyperopic PE in long eyes. Conclusions: The new HQST was superior to the original HQ and reached statistical and clinical results comparable to those achieved by the BUII, EVO, Kane and RBF formulas.