Background We evaluated a magnetic resonance (MR)-conditional high-power microwave ablation system. Methods An exvivo 1.5-T evaluation was conducted by varying the sequence (T1-weighted volume interpolated breath-hold examination, T1w-VIBE; T1-weighted fast low-angle shot, T1w-FLASH; T2-weighted turbo spin-echo, T2w-TSE), applicator angulation to B0 (A-to-B0), slice orientation, and encoding direction. Tip location error (TLE) and artefact diameters were measured, and influence of imaging parameters was assessed with analysis of variance and post hoc testing. Twenty-four exvivo ablations were conducted in three bovine livers at 80 W and 120 W. Ablation durations were 5, 10, and 15 min. Ablation zones were compared for short-axis diameter (SAD), volume, and sphericity index (SI) with unpaired t test. Results The artefact pattern was similar for all sequences. The shaft artefact (4.4 ± 2.9 mm, mean ± standard deviation) was dependent on the sequence (p = 0.012) and the A-to-B0 (p < 0.001); the largest shaft diameter was measured with T1w-FLASH (6.3 ± 3.4 mm) and with perpendicular A-to-B0 (6.7 ± 2.4 mm). The tip artefact (1.6 ± 0.7 mm) was dependent on A-to-B0 (p = 0.001); TLE was -2.6 ± 1.0 mm. Ablation results at the maximum setting (15 min, 120 W) were SAD = 42.0 ± 1.41 mm; volume = 56.78 ± 3.08 cm3, SI = 0.68 ± 0.05. In all ablations, SI ranged 0.68–0.75 with the smallest SI at 15 min and 120 W (p = 0.048). Conclusion The system produced sufficiently large ablation zones and the artefact was appropriate for MR-guided interventions.