Let
$$
f(x,y,z)=\log(5y+2)+\log(2z+5)+\log(x+3y)+\log(3x+z)-\log(xyz).
$$
Then
$$
f_x=\frac{-1}{x}+\frac{1}{x+3y}+\frac{3}{3x+z};\\
f_y=\frac{-1}{y}+\frac{3}{x+3y}+\frac{5}{2+5y};\\
f_z=\frac{-1}{z}+\frac{1}{3x+z}+\frac{2}{5+2z}.
$$
So the FOC's yield
$$
x=1,\quad y=\sqrt{2/15}, \quad z = \sqrt{15/2}.
$$
With these values, the original function evaluates to $241+44\sqrt{30}$. To be rigorous, you can verify the SOC's but I think it will probably work. According to Mathematica's FindMinimum, the values of $x$, $y$, and $z$ above solve the problem.
P.s. A commment on why naive AM-GM may not work: Suppose you split the 4 expressions according to
\begin{gather}
\frac{x}{m}+\cdots+\frac{x}{m}+\frac{3y}{n}+\cdots+\frac{3y}{n}\tag{i}\\
\frac{3x}{p}+\cdots+\frac{3x}{p}+\frac{z}{q}+\cdots+\frac{z}{q},\tag{ii}\\
\frac{5y}{r}+\cdots+\frac{5y}{r}+\frac{2}{s}+\cdots+\frac{2}{s},\tag{iii}\\
\frac{2z}{c}+\cdots+\frac{2z}{c}+\frac{5}{d}+\cdots+\frac{5}{d}.\tag{iv}
\end{gather}
The first 2 equations say
$$
x=\frac{3m}{n}y,\quad z=\frac{3q}{p}x.
$$
You want the exponent of $x$ after applying AM-GM to match the exponent of $x$ in the denominator, which is $1$, so $\frac{m}{m+n}+\frac{p}{p+q}=1$, which simplifies to $\frac{m}{n}=\frac{q}{p}$. So (i) and (ii) imply
$$
z=\frac{9m^2}{n^2}y.\tag{A}
$$
Next, (iii) implies $y=\frac{2r}{5s}$. And since we want $\frac{n}{n+m}+\frac{r}{r+s}=1$, it must be that $\frac{r}{s}=\frac{m}{n}$. So (iii) implies
$$
y=\frac{2}{5}\frac{m}{n}.\tag{B}
$$
Similarly, (iv) implies $z=\frac{5c}{2d}$ with $\frac{c}{d}=\frac{p}{q}$ due to $\frac{c}{c+d}+\frac{q}{p+q}=1$. But $\frac{p}{q}=\frac{n}{m}$, so (iv) implies
$$
z=\frac{5}{2}\frac{n}{m}.\tag{C}
$$
Putting together (A), (B), and (C), we see that
$$
9\frac{m^2}{n^2}=\frac{z}{y}=\frac{5n}{2m}/\frac{2m}{5n}\implies\frac{9m^2}{n^2}=\frac{25n^2}{4m^2}.
$$
Taking square roots and simplifying yield
$$
6m^2=5n^2.
$$
But now we have a problem because $\sqrt{\frac{5}{6}}$ cannot be rational.