Cultural difference pose challenges for machine translation as it’s challenging to find equivalent words to convey cultural concepts. Therefore, this research conducted to find out the accuracy level of two popular machine translation namely Chat-GPT and Google Translate in translating Culture-specific Items from the bilingual description of historical heritage at Museum Negeri Nusa Tenggara Barat. Using HTER by Snover (2006), researcher try to find the level of accuracy of the MT. After long analysis of the accuracy of MT namely Chat-GPT and GT using the HTER theory by Snover, researcher found that between Chat-GPT and GT almost play accurate with the presentation 99.94% and 99.93% while the error rate for Chat-GPT is 0.06% while GT is 0.07%. Furthermore, the quality index (TQI) of the MTs are in the same range, which is EXCELLENT.