Nevin Climenhaga and Daniel Rubio have an interesting paper arguing against Molinism. While I've often been very critical of Molinism myself, I don't think their argument works at all, although they make a number of valuable points in the course of their discussion. Their argument is:
(1) If Molinism is true, then there is some set of facts Γ that fully explains Eve’s sinning and everything Eve does that influences whether she sins.
(2) If Γ fully explains S’s φ-ing as well as everything S does that influences whether S φ's, then S does not φ freely.
Therefore, (3) If Molinism is true, Eve does not freely sin.
The problem with the argument is that (2) is obviously false, despite the efforts of Climenhaga and Rubio to argue that it should be accepted by advocates of free will. Part of the problem is that 'S freely chooses to phi' is itself a fact; therefore there is always some set of facts that fully explains S's phi-ing, as well as everything S does that influences S's action, because you can always have a set that includes the influences and S's free choice itself. (2) is not, contrary to the claim made in the paper, a libertarian premise; all libertarians hold that there is a set of facts that fully explains somebody's freely doing something, namely, the set of facts that gives the influences and their free choice to do it.
Moreover, Molinism itself can be seen to reject (2) outright. On the Molinist view, there is a divine middle knowledge between God's knowledge of all possible things and God's knowledge of all actual things; God not only knows what all things can do, He not only knows what all things do, He also knows what all things would do in any given situation. On the very traditional versions, such as those found in Molina himself, God knows this 'would' by supercomprehension; God knows your individual nature very, very well, so that He knows what you would do. This is related, I think, to the idea that God knows you not only in the abstract but also insofar as you could play a role in any particular order of nature that God might create. Later Molinists do a lot more handwaving, but my point is not about this issue in particular but the fact that it shows that Molinism is already a rejection of (2): All Molinists are committed to there being facts that fully explain your free action, by the very idea of middle knowledge. All the authors have done is beg the question against Molinism.
(It doesn't help their argument that they try to push (2) by arguing that it is superior to a more standard principle of alternative possibilities, based on Frankfurt cases. But as I've been arguing almost as long as I've had this blog, Frankfurt examples are all based on a sleight of hand in which a scenario is first described in terms of alternative possibilities and then re-described as not having any. Ironically, I think this is pretty much what Molinism itself does -- what it describes as middle knowledge is sometimes described in a way that would make it knowledge of possibles and sometimes in a way that would make it knowledge of actuals, and by oscillating back and forth the illusion of a middle ground is created, just as the double description in a Frankfurt example creates the illusion of a free choice with no alternative possibilities. Thus (2) is not, in fact, superior to the principle of alternative possibilities, but the point is irrelevant anyway, since no thinking Molinist would accept (2) in the first place.)