OBJECTIVES: Ultra-high dose rate irradiation (UHDR) has been shown to spare normal tissue in various model systems. This study evaluates its potential to sterilize cancer cells using spheroid tumor models. METHODS: Spheroids from glioblastoma (U87), hypopharyngeal squamous cell carcinoma (2 sizes, FaDusmall and FaDularge) and breast adenocarcinoma (T47D) cells were irradiated with electron beams using UHDR (>200 Gy/s) or conventional dose rate (CONV, ∼0.1 Gy/s) exposures under ambient or reduced oxygen (1%) conditions. U87 and FaDusmall were also irradiated with protons. Spheroids were monitored using imaging for up to 100 days to determine the dose required to cure 50% of spheroids (SCD50). These data were used to calculate dose-modifying factor estimates for UHDR at the 50% survival level (DMFSCD50). RESULTS: A total of 3230 spheroids were analyzed. Under ambient oxygen tension, UHDR and CONV showed no significant differences in U87 (DMFSCD50 = 0.98, P = .47), FaDusmall (DMFSCD50 = 1.01, P = .75), and T47D (DMFSCD50 = 1.04, P = .25), regardless of electron or proton irradiation. Under reduced oxygen levels, significantly higher UHDR doses were required to sterilize the spheroids, with DMFSCD50 1.14 (U87, P < .01), 1.07 (FaDusmall, P = .02), and 1.13 (T47D, P < .01). FaDularge-spheroids irradiated under ambient oxygen showed a DMFSCD50 of 1.66 (P < .001). CONCLUSION: Using spheroid tumor models with long follow-up, we demonstrate that efficacy of UHDR varies across cancer types and conditions. Whereas small spheroids exhibit iso-efficacy, both reduced oxygen tension and increased spheroid size lead to higher DMF. ADVANCES IN KNOWLEDGE: This preclinical study suggests that tumor iso-efficacy with UHDR may not hold true for all cancer types and is associated with oxygen level.