We have heaps of Plato, actually. Somewhere between thirty and forty dialogues survive. (See here, for example.)

We don't rely on Arabic translations for any of Plato or Aristotle. At all. They have virtually nothing to do with the survival of the text of any ancient Greek philosopher to the present day. They have a lot to do with the popularity of certain authors, but not their survival. Mediaeval Arabic and Persian commentaries and translations were an important avenue for mediaeval western Europeans to get access to Aristotle. In the modern era, however, we no longer require that avenue.

Modern editions of these and other ancient Greek authors are based entirely on copies made and preserved in the Greek-speaking world, that is to say the eastern Roman empire in the Byzantine era, which had a rich tradition of transmitting old books and writing new ones throughout the mediaeval period. This fact is sometimes neglected, unfortunately: westerners who deal with these texts sometimes treat mediaeval and modern Greek practices and traditions with scathing contempt. Rudolf Pfeiffer's History of classical scholarship (1968-1976), for example, says nothing at all about the role played by mediaeval and modern Greek scholars in the transmission, preservation, and publication of ancient literature.

If we did rely on Arabic versions for ancient Greek texts, we would be sadly impoverished. Arab scholars had no interest in Greek historiography or poetry, for example. The first Arabic translation of Homer wasn't written until the early 20th century.

I've written some pieces on this which may fill in some gaps: part 1 to dispel the myth that we rely on Arabic versions, part 2 to tell a more accurate story of how ancient Greek texts got transmitted to the present day.