I have many times seen Lagrangian written as $\mathcal{L}(x(t),x'(t),t)$. I understand this is a function of $x(t)$, $x'(t)$ and $t$.
So it can theoretically look something like
$$\mathcal{L}(x(t),x'(t),t) = 5\cdot x(t)+(x'(t))^2+2\cdot t$$
(and therefore $\mathcal{L}(a,b,c) = 5\cdot a+b^2+2\cdot c$)
But why don't we simply write
$$\mathcal{L}(x(t),t) = 5\cdot x(t)+(x'(t))^2+2\cdot t$$
(and therefore $\mathcal{L}(a(t),b) = 5\cdot a(t)+(a'(t))^2+2\cdot b$)
Is the $x'(t)$ there just to make it clear that Lagrangian has it or are there other reasons?
The reason is because what you are considering is not exactly the Lagrangian. The Lagrangian would usually be a function
$$\mathcal{L}:\mathbb{R}^3\supset\kern{-3px}\to\mathbb{R}, \quad (x_1,x_2,x_3)\mapsto\mathcal{L}(x_1,x_2,x_3).$$
When you then consider a functional on the form
$$J[y]=\int_a^b \mathcal{L}(x,y(x),y'(x))\,\mathrm{d}x,$$
you are actually first composing your Lagrangian with the function $x\mapsto(x,y(x),y'(x))$. If you view it like this, then it should be clear why $\mathcal{L}$ must depend on three variables, and not just two, as in the definition of the Lagrangian itself, $x_2$ and $x_3$ (which we later replace with $y(x)$ and $y'(x)$ under the composition) have nothing to do with each other.