Something close to three thousand works were attributed to Jabir, which was already suspicious to medieval Arabic bibliographers, Ibn al-Nadim, in the tenth century, records that some of his contemporaries doubted Jabir had existed at all.
The dating was settled philologically by Paul Kraus in a two-volume study published in Cairo in 1942-43. Kraus showed on internal evidence, technical terms not coined until the mid-ninth century, and citation of Greek works that had not been translated into Arabic when the supposed author died, that the corpus cannot be the work of a single eighth-century author, and belongs instead to a school writing over roughly a century between about 850 and 950. That conclusion has held. His further identification of the authors with Ismaili circles has not: the Shi'ism of the texts has since been shown to differ substantially from Isma'ilism, and it should not be repeated as current scholarship.
What remains genuinely open is narrower than it is usually made to sound. The dispute among specialists, Fuat Sezgin, Pierre Lory and Syed Nomanul Haq against the strong form of Kraus, is not about the corpus's overall date but about whether an authentic eighth-century core survives underneath the later redaction. The traditional birth year of about 721 is itself an inference from the date of his father's execution rather than a record.