Linear Models & Regression — prediction from the better-fitting least-squares model

For a bivariate data, using the principles of least-squares, the following models were constructed: E1: $Y = 3X + 5$ with $\sum(Y_i - 3X_i - 5)^{2} = 15.48$; E2: $Y = -X^{2} + 8X + 2$ with $\sum(Y_i + X_i^{2} - 8X_i - 2)^{2} = 13.36$. What is the estimated value of Y for X = 6?