This study aims to validate the simulation model of the GAMOS/ GEANT4 code for a 6 MV photon beam produced by the Elekta Synergy Agility linear accelerator installed at the International Cancer Center of Dakar (CICD), Senegal. The simulation encompasses all major components of the accelerator head: the target, primary collimator, flattening filter, ionization chamber, and X and Y jaws, using a homogeneous water phantom. The phase space was placed after the jaws, and for each angular distribution model studied: Tsai, Koch–Motz 2BS, and Koch–Motz 2BN, the dose distribution was evaluated. This includes depth dose curves for field sizes of 5 × 5 cm² and 10 × 10 cm² at a source-to-axis distance (SAD) of 100 cm, as well as dose profiles at depths of 5, 10, 15, and 20 cm in the phantom, with a source-to-surface distance (SSD) of 90 cm from the target. The three bremsstrahlung angular distribution models implemented in GAMOS were then compared with experimental measurements. Validation was performed using the gamma index, with an acceptance criterion of 3% for dose difference (DD) and 3 mm for distance to agreement (DTA). For the depth dose curves, a 94% agreement was observed between simulated and experimental data for the 5 × 5 cm² field, and 96% for the 10 × 10 cm² field, regardless of the model. Regarding the dose profiles, the three models: Koch–Motz 2BN, Koch–Motz 2BS, and Tsai, exhibit perfect agreement (100%) with measurements for the 5 × 5 cm² field size at all depths. For the 10 × 10 cm² field, the Koch–Motz 2BN model shows excellent agreement of 100% at 5 cm and 20 cm depths, followed by the Tsai model with 99% at 20 cm. At 10 cm depth, agreement reached 99% for Koch–Motz 2BN and 97% for Tsai. At 15 cm, Koch–Motz 2BN and Tsai achieved 98%, followed by Koch–Motz 2BS with 92%. At 20 cm, Koch–Motz 2BN maintained 100% agreement, followed by Tsai (99%) and Koch–Motz 2BS (94%). This study compares three bremsstrahlung angular distribution models in GAMOS with experimental values, assessing their respective performances in photon beam simulation. These results may guide radiotherapy practitioners in selecting the most appropriate model. In summary, this work contributes to the validation and enhancement of simulation techniques in radiotherapy, thereby improving treatment optimization and patient safety in cancer care.