Where did I go wrong with my mesh analysis equations? Dependent Current Source

Here's the circuit and mesh conventions I chose (with a hope that it was close enough to your choices):

(In the above, I've included your equations as I've tried to assign them to my loops. I don't agree with your equations if I got the loop assignments assigned correctly. But I kept them in the above image for clarity, not accuracy.)

The three mesh equations, and the dependent source equation, are then:

$$\begin{align*} 0\:\text{V}+v_{\small{4\, ix}}-\left(-1\,j\:\Omega\right)\cdot\left(i_1-i_3\right)-1\:\Omega\cdot\left(i_1-i_2\right)&=0\:\text{V}\tag{$i_1$} \\\\ 0\:\text{V}-1\:\Omega\cdot i_2-1\:\Omega\cdot\left(i_2-i_1\right)-4\:\text{V}&=0\:\text{V}\tag{$i_2$} \\\\ 0\:\text{V}+4\:\text{V}-\left(-1\,j\:\Omega\right)\cdot\left(i_3-i_1\right)-1\:\Omega\cdot i_3&=0\:\text{V}\tag{$i_3$} \\\\ i_1&=4\cdot i_2\tag{$4\,i_x$} \end{align*}$$

This solves out as: \$i_1= 8\:\text{A}\$, \$i_2= 2\:\text{A}\$, \$i_3= \left(6 - 2\,j\right)\:\text{A}\$, and \$v_{\small{4\, ix}}= \left(8 - 2\,j\right)\:\text{V}\$.

Assuming \$v_{o_-}=0\:\text{V}\$ then it follows that \$v_{o_+}=1\:\Omega\cdot\left(6 - 2\,j\right)\:\text{A}=\left(6 - 2\,j\right)\:\text{V}\$.

Which works out to:

abs(6 - 2*j).n()
6.32455532033676
(arg(6 - 2*j) * (360/(2*pi))).n()
-18.4349488229220

or \$\approx 6.325\:\text{V}\:\angle -18.435^\circ\$.

I'll let you examine my equations and mesh current assignments to spot differences.