This got so complicated that I made a spreadsheet.
Note that there are two references for elevation in NA and they can vary by a few feet, National Geodetic Vertical Datum of 1929 (NGVD29) and North American Vertical Datum 1988 (NAVD88). It is OK to compare elevations using the same vertical datum, but don't compare across them.
I also have elevations from the National Elevation Dataset. That appears to be the most recent. I downloaded LIDAR data for the area, but I have no idea what to do with a ".las" file. The best NED data is good to one meter. I'm not sure what the error is for the Mount Tamalpais ridge data.
The historical USGS topo gives elevations as of 1952, which is after the peak was bulldozed. All elevations in this spreadsheet are after that.
After all this, I believe that Lists of John might be wrong. Those elevations are from the most recent USGS map. The elevation for West Peak needs to be interpolated from the contour lines. The marker is on the fourth line above 2400 feet, one contour below 2600 feet. My guess is that someone misread the 40 foot contours as 20 foot contours and recorded 2580 (2600 - 20) instead of 2560 (2600 - 40). Easy mistake to make.
Peakbagger appears to get their elevation numbers from topo maps, but a one meter difference is almost certainly beyond the accuracy of the data without a dedicated survey.
Finally, the county highpointers site has some discussion about this. They are convinced that West Peak is higher. One report references a 2003 survey. There should be some survey data because of the nature restoration work on West Peak and because East Peak is a fire lookout.