This article explores how different interest rate models perform in low-rate environments, focusing on the risk of negative interest rates. The researchers use historical data to analyze two multi-factor models and find that the probability of negative rates is significant in one model but zero in another due to lognormality. By using Monte Carlo simulations, they compare prices of financial instruments and show that the models do not always produce the same prices, especially near current interest rates. Additionally, they demonstrate the absence of arbitrage in both models when pricing long butterfly spreads.