To get results with the highest biological relevance, I would recommend to use primary hepatocytes. That will increase the value of your work and you can publish in higher ranking journals. We have good experience with the cells from Lonza: https://bioscience.lonza.com/lonza_bs/US/en/hepatocytes-for-drug-metabolism-studies?gclid=Cj0KCQiA_rfvBRCPARIsANlV66Ogh0siHKOgulHHVGRSplzp14BbS5IUv-iGXwCj9mWjdIGSAuhn_RoaAjXoEALw_wcB
A cell line which has been shown to be very authentical is the HepRG cell line (Lonza).: https://bioscience.lonza.com/lonza_bs/US/en/Primary-and-Stem-Cells/p/000000000000222176/NoSpin-HepaRG-%288-0M-cells-vial%29
I would concur with the suggestion that Peter has made. Primary cells are still considered more representative than hepatoma- or hepatocellular carcinoma-derived cell lines (HepG2 and C3A sub-categories are widely used for hepatoxicity investigations and metabolism studies) but there are more ethical implications involved. However, I would suggest the answer to your question is dependant on the specific species you are interested in studying. The AML12 cell line is a normal mouse-derived hepatocyte that may be useful.
It would depend on your end application. HepG2 is a commonly used hepatic cell line. It has certain functionalities of hepatic cells. In addition to HepG2 HepaRG and Can10 cell lines are commonly used. These cell lines are better for studying bile canaliculi functionalities.