I've found two type-settings of the prelude:
The original manuscript/facsimile is also available via IMSLP.
I'm not sure if I can trust either typeset versions. While a single A♭ in these editions, per the manuscript the second beat in the treble clef of measure six is clearly a G♭-B♭ chord.
The manuscript is incomplete, error-prone, and difficult to read. Are there any reasons to assume the typeset editions are correct?
A picture is worth a thousand words. Measure 6: typeset on the left, manuscript on the right. The manuscript's sixth measure excludes a would-be crossed-out measure 5. Notice that the manuscript's tie seems to contradict the G♭-B♭ chord. While A♭ is the better choice harmonically, melodically the G♭-B♭ chord isn't too out-of-place, and sounds more "haunting". Rhythmically
G♭-G♭-F, E♭-E♭-D♭ is more symmetrical than
A♭-G♭-F, E♭-E♭-D♭. Also, there's no way the G♭-B♭ chord is a typo; it was explicitly written with the octave D♭ in the bass clef as the 2nd inversion of the IV chord.