Usability is important for these lenses.
A faster zoom lens is more valuable, but a shorter zoom range is not.
The 28-70mm f/2 came for about 1k more than the RF 24-70mm f/2.8 (yes, I know the 28-70 came first). All it sacrificed was 4mm on the wide end, so the trade-off was minor.
If it's a standard zoom, my guess is Canon won't make it unless it covers at least three main focal lengths, like the 28-70 does (28, 35, 50) and enters wide angle territory (so no 35-85mm, for instance). 50 could be rounded from 45, I think- I don't believe such a small difference would discourage anyone.
If we're talking halo lens, they want to break new ground. The 28-70mm f/2 wasn't the first f/2 full-frame zoom, that title belongs to the Sigma 24-35mm f/2 Art, and I doubt this time they would accept matching apertures that have been done before, so it would have to be faster than f/1.8, because the Sigma 28-45mm f/1.8 exists.
24 or 28 to 45 or 50 at f/1.4, for 1k more than the 28-70mm f/2? Could work, I guess.
20-40, 24-45, 24-50, 28-45, 28-50 at f/1.4...I think those would be good possibilities.
EDIT: I realised now pretty much all these were patented. I wasn't looking at that.