To evaluate the possible range of application of the new InCise2 MLC for the CyberKnife M6 system in brain radiosurgery, a plan comparison was made for 10 brain metastases sized between 1.5 and 9cm3 in 10 patients treated in a single fraction each. The target volumes consist of a PTV derived by expanding the GTV by 1mm and were chosen to have diversity in the cohort regarding regularity of shape, location and the structures needed to be blocked for beam transmission in the vicinity. For each case, two treatment plans were optimized: one using the MLC and one using the IRIS-collimator providing variable circular fields. Plan re-quirements were: dose prescription to the 70% isodose line (18 or 20Gy), 100% GTV coverage, ≥98% PTV coverage, undisturbed central high dose region (95% of maximum dose) and a conformity index as low as possible. Plan com-parison parameters were: conformity index (CI), high-dose gradient index (GIH), low-dose gradient index (GIL), total number of monitor units (MU) and expected treatment time (TT). For all cases, clinically acceptable plans could be gen-erated with the following results (mean±SD) for CI, GIH, GIL, MU and TT, respectively for the MLC plans: 1.09±0.03, 2.77±0.26, 2.61±0.08, 4514±830MU and 27±5min and for the IRIS plans: 1.05±0.01, 3.00±0.35, 2.46±0.08, 8557±1335MU and 42±7min. In summary, the MLC plans were on average less conformal and had a shallower dose gradient in the low dose region, but a steeper dose gradient in the high dose region. This is accompanied by a smaller vol-ume receiving 10Gy. A plan by plan comparison shows that usage of the MLC can spare about one half of the MUs and one third of treatment time. From these experiences and results suggestions for MLC planning strategy can be de-duced.