A large-scale stated preference survey and advanced two-stage behavioral model designed to quantify how travelers respond to congestion pricing. The survey integrated real-world travel data from Google APIs into each respondent's choice scenarios and used a multi-stage choice design that mirrors real decision-making — capturing detailed within-mode preferences across car, public transport, and cycling. The behavioral model combines a Scaled Multinomial Logit (exploiting the full survey structure) with a Nested Logit for reliable real-world forecasting, validated with exceptionally high accuracy across all population segments. The findings provide the empirical foundation for congestion pricing scheme design, optimal pricing, and the complementary transit investments needed to maximize effectiveness.