A fundamental problem of this question is the a priori of "the current consensus of the writing of the gospels". Such a consensus does not exist. This two-sources-theory is just one of several attempts to reconcile the difficulties in explaining the synoptic problem, and admitted: fairly popular at that.
There are other theories, which like this one have their advantages and disadvantages. These may be grouped into Markan priority, Matthean priority, Lukan priority and a multitude of sources. An overview of the most common models is found at Wikipedia and at the Synoptic Problem Website. Alas, those fail to mention a tradition of explanation that goes from Harnack to Klinghardt: the (proto-)Marcionite priority, which complicates matters further. (I am aware that the last line is extremely unpopular in America. It has its merits, but is less useful here since that passage seems to be absent in the reconstructed gospel.)
That said, the passage in question is indeed very problematic for most theories. One example from the literature is Zeba Antonin Crook, "The Synoptic Parables of the Mustard Seed and the Leaven: A Test-Case for the Two-Document, Two-Gospel, and Farrer-Goulder Hypotheses", Journal for the Study of the New Testament 22.78 (2000): 23-48 —
As long as the various source hypotheses remain hypotheses, the synoptic problem shall remain unsolved and even unsolvable. The nature of any hypothesis is that it cannot account for all of the data all of the time, for in so doing it would cease to be a hypothesis. The data presented by the synoptic parables of the Mustard Seed (Mt. 13.31-32//Mk 4.30-32//Lk. 13. I 8-19) and the Leaven (Mt. 13.33//Lk. 13.20-21) afford us with a valuable opportunity to test the claims of each of the three current and popular solutions to the synoptic problem – the 2 Document Hypothesis (2DH), the Two-Gospel Hypothesis (2GH) and the Farrer-Goulder Hypothesis (F-G).
It is not uncommon to read studies that either state explicitly or work under the assumption that the synoptic problem has been solved. Using the parables of the Mustard Seed and the Leaven as a test-case, it becomes clear that the problem is far from a solution. Each of the three major source hypotheses has its strengths (and weaknesses) when it attempts to account for the data generated by these two pericopae. Although this paper concludes that the Two-Document Hypothesis (2DH) deals with the data with the fewest problems, the strengths of the other hypotheses coupled with the weaknesses of the 2DH should help keep the 2DH honest.
It is therefore currently not possible to reach a definite conclusion. Whole dissertations try to solve the problems with this passage, for example: Franz Kogler: Das Doppelgleichnis vom Senfkorn und vom Sauerteig in seiner traditionsgeschichtlichen Entwicklung, Echter Verlag: Würzburg, 1988.
Kogler delivered an answer, but that met criticism as well (Bartosz Adamczewski, Q Or Not Q?: The So-called Triple, Double, and Single Traditions in the Synoptic Gospels [Peter Lang, 2010], p. 121):
… deutero-Mark transformed the Markan parable of the mustard seed into a twin parable of the mustard seed and of the yeast by adapting it to the hellenistic cultural milieu […]
Kogler failed to adresss adequately the redaction-critical issue of correct attribution of the composition of the twin, gender-paired parables within the synoptic tradition…
(It might also be noted that Adamczewski's work was also heavily criticized.) If we follow Crook's assessment in comparing three theories in their explanatory power for this pericope – alone – than it is indeed a 'yes, it is plausible' to the question: according to the two document hypothesis:
This makes it reasonable to think that Luke found it in Q and left it there.
But this is just for this pericope and only comparing three of the popular theories. For the whole synoptic problem the findings presented are much more limited in problem solving and should only be read as an indicator, not as proof for the whole theory.
These are the passages from Crook that mainly praise the 2DH, while highlighting only a few of its shortcomings:
That a pericope is triple tradition is generally sufficient to show that Mark is the only source for Matthew and Luke. The Mustard Seed presents certain problems for this. While the degree and nature of the agreements between Matthew and Mark (see Table 1) demand direct literary dependence, the agreements between Mark and Luke do not. There are several features that suggest Lukan independence of Mark here. The words Luke shares with Mark are weak; they do not add anything significant to the meaning or application of the parable. Additionally, the only material which shows that Luke and Mark are even telling the same story is the triple agreement material; therefore, there is nothing to demand Luke’s reliance on Mark.
Further, all the double agreement words of Luke/Mark against Matthew are comfortable enough in Luke that there is no need to suggest Mark’s influence at all. The τινι (v. 18a) which Luke appears to take from Mark is actually part of a Lukan construction (τινι ομοισς) which Luke uses without Markan parallels (Lk. 6.47; 7.30; 13.20). This suggests that Luke need not have obtained this word (even as minor as it is) from Mark. The unnecessary order of the two parables also supports the contention that two versions of the Mustard Seed survive in our text. Therefore, according to the 2DH, two different versions of this parable are preserved in the synoptic Gospels––that of Mark and Q.
[…] Tuckett suggests (for the Mustard Seed) that while Mark was working from an earlier source, our parable offers no reasons to assume that Q itself (in the form available to us) is that source. The basic problem with proposing too many Mark/Q overlaps is the nature of the evidence; the data that alerts us to the possibility of a Mark/Q overlap, namely that Mark shares so little in common with Matthew/Luke(Q), is the same data that militates against Mark’s relying directly on Q.
On the 2DH, Matthew conflates Mark and Q here as he is said to do
consistently. Matthew is required to have made certain alterations, some of which are more easily rationalized than others.[…]
Next, we need to look at how Luke deals with the two versions of the Mustard Seed. According to the 2DH, Luke prefers Q and alters it very little when confronted by a similar passage in Mark. While neither Matthew nor Luke has the monopoly on original wording concerning Most 2DH commentators agree that Luke reflects the Q version of the Mustard Seed more closely than Matthew. Although Luke follows Q very closely, there are a few changes that Luke likely made to Q.[…] There is considerable debate on whether Luke’s double opening question reflects Markan influence.25 It was noted earlier that since Luke is five chapters beyond Mark’s Mustard Seed, the 2DH can easily account for the meager agreements between the two versions. Luke’s double question presents the 2DH with a problem, however, since it cannot be said that he categorically avoids this characteristic feature of Mark. It is not fair to claim that Luke is too distant from Mark to take any significant vocabulary, or claim, as does Friedrichsen. […]
The 2DH must decide whether Luke is using Mark for the Mustard Seed or not. Fleddermann claims that both of Luke’s double questions contain similarly un-Lukan language, but this cannot tell us Luke’s source.[…]
There are three problematic words in Luke which the 2DH sidesteps
by suggesting Luke found them in Q. If this is not entirely satisfactory, it is a far better hypothesis than could be suggested by the 2GH or F-G
hypothesis who, as we shall see, must ascribe them to Lukan redaction. These words are κηπος and βαλλω in the Mustard Seed and παλιν in the Leaven. There is little to indicate that Luke ’liked’ κηπος and hence may have been predisposed to add it to his source, since it occurs only here in Luke (and never in Mark or Matthew). This makes it reasonable to think that Luke found it in Q and left it there. The image of impurity that is created by planting mustard in a garden is mirrored in the Leaven, and thus might have been a part of the original radical parables.
So, while it remains impossible to account for why Matthew removed παλιν from the Leaven, its occurrence in Luke’s Leaven is best explained by his finding it in Q and keeping it as he had before.
To review briefly, on the whole, Matthew’s conflation of Mark and Q is problem free, likewise with the majority of the changes he is observed to have made. The area where the 2DH shows weakness in its account of the Mustard Seed and Leaven is primarily in the double question in Luke, since for all other differences he is not looking at Mark. With respect to the historic presents in Matthew, his choice of αυξανω over αναβαινω, his avoidance of Mark’s επι της γης, and the inclusion of ελθειν of the birds, none of these contradict any claimed patterns of redaction, they merely cannot be explained without simply restating the problem. […]
We have seen in the course of this study that the three popular and
current source hypotheses can account for some of the data quite
adequately. This is not surprising; if they were not good to a certain degree, we should be surprised that they had any adherents at all. Not one of them was able to account for all the data equally well; some encountered problems dealing with specific words, others with presenting coherent redactional procedures. What separates the source hypotheses is not that one of them can answer all the questions posed to it, but that one has the fewest problems. It might be a sad state of affairs that we need to accept as the better answer the one that fails the fewest times, but this is the current state of the synoptic problem. For the Mustard Seed and the Leaven, the Two-Document Hypothesis deals with the data in the least problematic manner.