PurposeTo biomechanically evaluate a flat posterior cruciate ligament (PCL) reconstruction utilizing rectangular femoral bone tunnels.MethodsEight fresh‐frozen human knee specimens were tested in a six‐degrees‐of‐freedom robotic test setup. In each testing step, a force‐controlled test protocol was performed, including 89 N posterior tibial translation (PTT) in neutral, internal and external rotation, from 0 to 90° of flexion. After determining the native knee kinematics, the PCL was cut. Subsequently, a flat PCL reconstruction (PCLR) with a rectangular bone tunnel was performed, utilizing a quadriceps tendon autograft with a patellar bone block. After filling the bone tunnel, a single‐bundle PCLR without and with femoral interference screw fixation, as well as a double‐bundle reconstruction, was performed. Statistical analysis was performed using mixed linear models.ResultsCutting of the PCL led to significant (p ≤ .05) increases in PTT, from 0 to 90° of flexion, up to 10.7 mm, in comparison to the native state. After flat reconstruction and double‐bundle reconstruction, no significant difference was found between the native and reconstructed state (p ≥ .05). The single‐bundle PCLR without interference screw showed significantly increased PTT in comparison to the native state in 30° (mean difference [MD] 3.3 mm; 95% confidence interval [CI] 1.3 – 5.2 mm; p < .001), 60° (MD 4.4 mm; 95% CI 2.5–6.4 mm; p < .001) and 90° of flexion (MD 4.0 mm; 95% CI 2.1–6.0 mm; p < .001). The single‐bundle PCLR with additional interference screw showed significantly increased PTT in comparison to the native state only in 30° (MD 1.9 mm; 95% CI 0.05–3.8 mm; p = .01).ConclusionBoth a flat and a double‐bundle PCLR were able to restore the native knee kinematics in all tested flexion angles. A single‐bundle reconstruction was not able to fully restore native kinematics, with only small residual anteroposterior instability.Level of EvidenceNot applicable (an experimental laboratory study).