Six tanh Layers and a Shrinking Gradient

A network stacks six identical layers. For $k = 1, \dots, 6$, layer $k$ computes

$a_k = \tanh(z_k), \qquad z_k = w\,a_{k-1}, \qquad w = 1.25$

and the same weight $w$ is shared by every layer. During one backward pass it happens that $\tanh(z_k) = 0.6$ at every layer. Recall that

$\frac{d}{dz}\tanh(z) = 1 - \tanh^2(z)$

The gradient arriving from the loss at the top of the stack is $\dfrac{\partial L}{\partial a_6} = 5$.

What is $\dfrac{\partial L}{\partial a_0}$?