Full disclosure, I'm a cancer research guy, not infectious disease. But certainly many of the animal models for human infectious diseases are fraught with limitations just like those in oncology.
The Gilead drug remdesivir, for instance, looked good in vitro for Ebola, and pretty good in primate disease models models - but perhaps has not shown as much promise in the field. The primate models are so-called semi-permissive for Ebola (and also some respiratory viruses), meaning that the disease often doesn't quite take hold as well in monkeys or is not as sustained as it is in humans. So, testing of the drugs in these models often is early on after exposure to the infectious agent to show if it has an impact.
Compare that to treating people in the field where Ebola has sprung up - people in all the various stages of disease, long after being infected with various levels of disease burden and the resulting complications of other systems involved. It's not so simple to make the leap from preclinical to clinical.
It all highlights the need for the continued clinical trials of all of these agents alone or in combination, as is ongoing now. Sporadic reports of a few patients treated here or there is not conclusive evidence of efficacy.