PurposeDetector‐dependent interinstitutional variations of the beam data may lead to uncertainties of the delivered dose to patients. Here we evaluated the inter‐unit variability of the flattened and flattening filter‐free (FFF) beam data of multiple TrueBeam (Varian Medical Systems) linear accelerators focusing on the small‐field dosimetry.MethodsThe beam data of 6‐ and 10‐MV photon beams with and without flattening filter measured for modeling of an iPLAN treatment planning system (BrainLAB) were collected from 12 institutions — ten HD120 Multileaf Collimator (MLC) and two Millennium120 MLC. Percent‐depth dose (PDD), off‐center ratio (OCR), and detector output factors (OF det) measured with different detectors were evaluated. To investigate the detector‐associated effects, we evaluated the inter‐unit variations of the OF det before and after having applied the output correction factors provided by the International Atomic Energy Agency (IAEA) Technical Reports Series no. 483.ResultsPDD measured with a field size of 5 × 5 mm2 showed that the data measured using an ionization chamber had variations exceeding 1% from the median values. The maximum difference from median value was 2.87% for 10 MV photon beam. The maximum variations of the penumbra width for OCR with 10 × 10 mm2 field size were 0.97 mm. The OF det showed large variations exceeding 15% for a field size of 5 × 5 mm2. When the output correction factors were applied to the OF det, the variations were greatly reduced. The relative difference of almost all field output factors were within ± 5% from the median field output factors.ConclusionIn this study, the inter‐unit variability of small‐field dosimetry was evaluated for TrueBeam linear accelerators. The variations were large at a field size of 5 × 5 mm2, and most occurred in a detector‐dependent manner.