Since the 1970s, many biofidelity targets for the thorax in frontal impact have been proposed for both dummy and finite element (FE) model application. This paper proposes an update using three test configurations:​ impactor, sled with a 3‐point belt restraint and table top. For the impactor configuration, an extensive Post Mortem Human Subject (PMHS) test database was constituted using both recent and old tests. The considered responses were the impactor force versus thoracic external deflection. The normalization process was updated using a mass‐spring model. Recent published methods were used to exclude outliers and pool the PMHS tests into consistent samples. For the sled configuration, a recent test series of 8 PMHS subjected to the same impact conditions was chosen. The considered responses were the shoulder belt tension versus time and the deflection versus time of 5 points at the anterior wall of the ribcage in the X, Y and Z directions. The responses were normalized using a mass‐based technique. Interpolation of missing portions of signals were performed for some tests. The T8 vertebra was chosen to define the thoracic coordinate system. The average PMHS response curve plus and minus one standard deviation was proposed as a biofidelity target for both the impactor and sled configurations. The table top configuration was considered in order to develop a biofidelity target for a distributed loading condition. A recent test series that used both distributed and diagonal belt loading conditions was chosen. To account for the limitations of this test configuration with respect to a real car crash, only the force values at 20% of thoracic compression were taken and were used relatively to the diagonal belt loading.