Formulaic sequences (FS), combinations of words that the natives memorize and use ho-listically to express a particular idea in certain communicative contexts, and that can be more than half of the speech, stand out in the lexical sub competence, central element of the development of communicative competence. The lexical approach determines that the mastery of these FS (chunks) has a positive effect on the improvement of the fluency in the communicative interaction of the learners. For Spanish as a foreign or second lan-guage (SFL), the Plan Curricular del Instituto Cervantes (PCIC) offers a non-exhaustive selection of these FS as an example of lexical units in each of the reference levels (A1-C2). For their selection, the criteria are frequency and communicative profitability, but it is done through an intuitive assessment based on the teaching experience. In this re-search, we set out to confirm the adequacy of this intuitive leveling, analyzing the fre-quency of FS of level B2 of the PCIC in various corpus available at present. Likewise, in an exploratory way, we will expose the usefulness of these corpus for data-driving learn-ing (DDL), with real examples of these FS of the target language.