Recent cases of severe toxicity during clinical trials have been associated with antiviral ribonucleoside analogs (e.g. INX-08189 and balapiravir). Some have hypothesized that the active metabolites of toxic ribonucleoside analogs, the triphosphate forms, inadvertently target human mitochondrial RNA polymerase (POLRMT), thus inhibiting mitochondrial RNA transcription and protein synthesis. Others have proposed that the prodrug moiety released from the ribonucleoside analogs might instead cause toxicity. Here, we report the mitochondrial effects of several clinically relevant and structurally diverse ribonucleoside analogs including NITD-008, T-705 (favipiravir), R1479 (parent nucleoside of balapiravir), PSI-7851 (sofosbuvir), and INX-08189 (BMS-986094). We found that efficient substrates and chain terminators of POLRMT, such as the nucleoside triphosphate forms of R1479, NITD-008, and INX-08189, are likely to cause mitochondrial toxicity in cells, while weaker chain terminators and inhibitors of POLRMT such as T-705 ribonucleoside triphosphate do not elicit strong in vitro mitochondrial effects. Within a fixed 3’-deoxy or 2′-C-methyl ribose scaffold, changing the base moiety of nucleotides did not strongly affect their inhibition constant (Ki) against POLRMT. By swapping the nucleoside and prodrug moieties of PSI-7851 and INX-08189, we demonstrated that the cell-based toxicity of INX-08189 is mainly caused by the nucleoside component of the molecule. Taken together, these results show that diverse 2′ or 4′ mono-substituted ribonucleoside scaffolds cause mitochondrial toxicity. Given the unpredictable structure-activity relationship of this ribonucleoside liability, we propose a rapid and systematic in vitro screen combining cell-based and biochemical assays to identify the early potential for mitochondrial toxicity.