merge sort algorithm time complexity is $\boldsymbol{O( n log n)}$ in all cases.
in first case ,for given input size 64 it is taking 30 secons , so $\boldsymbol{( n log _2n)}$. $C_1$=30 seconds
64 $log_2$ 64 $C_1$ =30 seconds
$C_1$ =$\frac{30}{64*6}$
for 6 minute ,input size will be- $\boldsymbol{( n log _2n)}$ $C_1$=6*60 seconds
$\boldsymbol{( n log _2n)}$=$ \frac {360} { C_1}$ ,put the value of $C_1$
$\boldsymbol{( n log _2n)}$=$ \frac {360} {\frac{30}{64*6}}$
$\boldsymbol{( n log _2n)}$= 4608
$\boldsymbol{( n log _2n)}$=$\boldsymbol{( 512 log _2512)}$
we can compare both side and $\boldsymbol n=512$.