Wolfram Alpha evaluates this integral numerically as
$$\int_0^{\infty} \frac{x^2}{\cosh^2 (x^2)} dx=0.379064 \dots$$
Its value is apparently
$$\frac{\sqrt{2}-2}{4} \sqrt{\pi}~ \zeta \left( \frac{1}{2} \right)=0.37906401072\dots$$
How would you solve this integral?
Obviously, we can make a substitution $t=x^2$
\begin{align} \int_0^{\infty} \frac{x^2}{\cosh^2 (x^2)} dx&=\frac{1}{2} \int_0^{\infty} \frac{\sqrt{t}}{\cosh^2 (t)} dt\\[10pt] &=\int_0^{\infty} \frac{\sqrt{t}}{\cosh (2t)+1} dt\\[10pt] &=\frac{1}{2 \sqrt{2}}\int_0^{\infty} \frac{\sqrt{u}}{\cosh (u)+1} du \end{align}
We could use geometric series since $\cosh (u) \geq 1$, but I don't know how it will help.
$$I=\frac{1}{2\sqrt{2}}\int_{0}^{+\infty}\frac{\sqrt{u}\,du}{1+\cosh(u)}=\frac{1}{\sqrt{2}}\int_{1}^{+\infty}\frac{\sqrt{\log v}}{(v+1)^2}\,dv=\frac{1}{\sqrt{2}}\int_{0}^{1}\frac{\sqrt{-\log v}}{(1+v)^2}\,dv \tag{1}$$ but since $$ \int_{0}^{1}v^k \sqrt{-\log v}\,dv = \frac{\sqrt{\pi}}{2(1+k)^{3/2}} \tag{2}$$ by expanding $\frac{1}{(1+v)^2}$ as a Taylor series we get:
and the claim follows from the well-known: $$ \eta(s) = (1-2^{1-s})\,\zeta(s)\tag{4} $$ that gives an analytic continuation for the $\zeta$ function.