So, when the coefficients are somewhat arbitary, I'm quite confused. Obviously there are 3 fixed points. There are other quite obvious ones, by letting $\lambda x+ \beta x^3 = -x$.
However when we look at its graph versus $y=x$, it seems for some choice of $\lambda$ and $\beta$, we can find a point that is greater than the positive fixed point, go between $0$ and the positive fixed point, and go back to this point by graphical analysis. The period is 2 (if possible, I just look at the graph and made an educated guess). Seems higher period is possible too if there are more steps between $0$ and positive fixed point. But I don't know if it's possible and how to restrict $\lambda$ and $\beta$.
The most straight forward idea is to compute $F^n$. However $F$ is already cubic and the power immediately goes out of control.
By the period-3 theorem, if you can find a period-3 point, then there are fixed points cycles of any period. So you only need to explore the roots of $(F^3(x)−x)/(F(x)−x)$ of $F(x)=ax-x^3$.
In the CAS Magma (online calculator) this can be computed as
To find the parameters for which the number of real roots changes, compute the discriminant relative to the variable $x$. As these roots come in multiples, of 3 for the period-3 points, the difficulty of the polynomial can be reduced by computing its factorization.
The result of this small script is
so that the relevant roots for the given task are at $2.4504409644878751728..$ and $2.9817718759600182561..$. Plotting $G$ around these values for $a=2.4, 2.5, 2.95, 3.05$ gives the graphs
which demonstrate that each of the 2 transitions introduces 4 period-3 cycles