Wellek is not ascribing any particular conceptual or philosophical view of realism to the Romantics in the passage quoted. Wellek says that in his essay:
I shall be content with raising the question of realism to the nineteenth century, anchored in a particular moment of history, referable to a well-known body of texts. ... I shall make some common sense distinctions and lead slowly to a concrete description of the period concept of realism, which I shall regard as a regulative concept, a system of norms dominating a specific time, whose rise and eventual decline it would be possible to trace and which we can set clearly apart from the norms of the periods that precede and follow it.
Wellek's definition of realism has to do with the qualities of literature in the middle and late nineteenth century: for example, the novels of Dumas fils, Balzac, Champfleury, Thackeray, and Eliot. He quotes David Masson's British Novelists and their Styles (1859), which calls Thackeray "a novelist of what is called the Real school" in contrast to Dickens, "a novelist of the Ideal or Romantic school". Wellek is specifically arguing that during this period, realism emerges as regulative and normative. That is to say, writers of the period adopted realism both as a term and as a norm to describe and regulate their writings. In other words, Wellek is examining and explaining "realism" as a specific technical term used in the history of literature, what he calls a period term.
Wellek contrasts his practice with two critical practices he finds deplorable:
one, the extreme nominalism which considers [period terms] mere arbitrary linguistic labels, a tradition prevalent in English and American scholarship, and the other, very common in Germany, of considering such terms as almost metaphysical entities whose essences can be known only by intuition.
This question makes the second error. As framed, the question assumes that "realism" is an entity outside geography and history—that it has a meaning equally applicable to all literary periods everywhere. When it asks "Was romanticism ambiguous to realism?" without defining those terms, it asks about the relationship between two "metaphysical entities whose essences are" intuitively obvious. To quote Wellek, of all critics, while framing the question in this way compounds the error, because his entire essay refutes such framing.
Wellek quotes Schiller, Schelling, Schlegel, and other Romantic or pre-Romantic writers in the section of the essay that describes the pre-history of the term "realism" in literary discourse, prior to its becoming a period term in the mid nineteenth century. The point of those quotations is to show the opposite of what your question assumes: Wellek is seeking to demonstrate that for those writers, realism did not have a fixed meaning. Or to be more precise, they did not use realism in a normative and regulatory way, they used it only ad hoc. It is therefore not surprising that their usage appears to pull in all directions. This isn't because they were "ambiguous toward realism". It's because the term "realism" did not mean something specific to them. They did not have a shared view toward realism, because realism didn't exist yet.
So if Schiller says that realism cannot make a poet, and Schlegel says that there is no true realism except that of poetry, Wellek's purpose in juxtaposing these two statements is to show that "realism" did not have a single definition in literary discourse prior to its adoption as a normative term by later nineteenth century writers. From that, one cannot conclude that Romantics were conflicted about their attitude toward realism, or indeed, anything about the Romantics' attitude toward realism, because on Wellek's terms there was no such thing as realism during the Romantic era.
Realism seems like such a transparent, intuitively grasped term to us, it feels that the claim of its historic specificity must be wrong. But it's actually the assumption that the word means the same thing across time and space that is mistaken. To ask about the Romantic attitude toward realism is akin to asking whether the 300 x 50 x 30 dimensions of Noah's Ark represents the number of units of quantum information used in its construction. Cubits and qubits are not the same thing, and applying the more recent term to the earlier era is a category error.
