Trackability of thrombectomy catheters through tortuous cerebral vessels is a key determinant of mechanical thrombectomy success, particularly for large-bore aspiration catheters. Yet, the underlying biomechanical challenges remain unclear. This study integrates in silico and in vitro analyses to investigate catheter navigation in a flexible intracranial vasculature model. A silicone patient-averaged tortuous vessel model was used for experimental studies in a circulatory flow loop and reconstructed from CT imaging for computational simulations. Regarding the in silico part of the study, in contrast to prior work relying on tip-dragging or centerline-based advancement, we implemented clinically realistic catheter pushing mechanics. We varied the vessel compliance and catheter-vessel friction coefficients to understand their sensitivity toward navigation. Strong qualitative agreement emerged between simulated and experimental catheter paths. Key findings include: (i) realistic pushing produced trajectories distinct from tip-dragging, with the catheter naturally aligning along the outer curvature to generate supportive contact and it matches with in vitro experiments; (ii) increased vessel flexibility (2 MPa) markedly improved catheter advancement, whereas stiffer vessels (10 MPa and rigid) promoted kinking; (iii) catheter-vessel interaction was observed to be a critical factor in navigation, with low friction coefficient (F) enhancing trackability (F < 0.1) and high friction (F > 0.15) triggering bending and kinking. Incorporating vessel flexibility and clinically representative pushing mechanics is essential for accurate thrombectomy modeling. The presented framework accurately reproduces catheter behavior, particularly in curved segments, and offers predictive capabilities for device performance. These insights offer quantitative design guidance for next-generation microcatheters and aspiration catheters, highlighting the critical role of catheter-vessel mechanics in distal cerebral arteries.