We consider the problem of broadcasting in an unknown radio network modeled as a directed graph $G=(V,E)$, where $|V|=n$. In unknown networks, every node knows only its own label, while it is unaware of any other parameter of the network, including its neighborhood and even any upper bound on the number of nodes. We show an $\mathcal{O}(n\log n\log\log n)$ upper bound on the time complexity of deterministic broadcasting. This is an improvement over the currently best upper bound $\mathcal{O}(n\log^2n)$ for arbitrary networks, thus shrinking exponentially the existing gap between the lower bound $\Omega(n\log n)$ and the upper bound from $\mathcal{O}(\log n)$ to $\mathcal{O}(\log\log n)$.