The tide-free release of the EGM2008 combined global geopotential model and its tide-free pre-release PGM2007A are comparedwith Australian land, marine and airborne gravity observations, co-located GPS-levelling on the [admittedly problematic] Australian Height Datum, astrogeodetic deflections of the vertical, and the AUSGeoid98 regional gravimetric quasigeoid model. In all comparisons, EGM2008 performs better than an y previous global gravity model. The standard deviation of the differences between free-air gravity anomalies from EGM2008 and free-air gravity anomalies from Australian land gravity ob-servations is ±5.5 mGal, compared to, e.g., ±11.7 m Gal for EGM96. Furthermore, the standard deviation of the differences between height anomalies from EGM2008 and a nation-wide set of 254 GPS-levelling points is ±17.3 cm, compared to, e.g., ±33.4 cm for EGM96. In the comparisons with GPS-levelling, EGM2008 also outperforms AUSGeoid98 (standard deviation of ±19.1 cm in the differences with the nation-wide set of 254 GPS-levelling points), and the same holds for the comparison to astrogeodetic deflections of the vertical. However, due to the poor quality of some of the Australian data, we cannot legitimately claim to truly validate EGM2008. Instead, EGM2008 confirms the already-known problems with the Australian data, as well as revealing some previously unknown problems. If one wants to claim validation, then EGM2008 is validated implicitly because it can confirm the errors in our regional data. Simply, EGM2008 is a good model over Australia.