A look at the cal kit coefficients of the 85032B on the Keysight website

shows for any VNA, the coefficients of the male open to be C0=62.14, C1=-143.07, C2=82.92, C3=0.76. Yet in my 8753ES, with the latest firmware (7.74), the capacitance coefficients are C0=74.576 C1=-914.38 C2=115, C3=-1.3537

Is this difference intentional? I did wonder these coefficients were optimized for a more limited frequency range of 3 GHz, as this is a 3 GHz 8753ES. But on hearing from others with 6 GHz instruments, those too have C0=74.576 C1=-914.38 C2=115, C3=-1.3537, so that does not explain this.

The offset delay of this standard, is 17.411 ps, both in the firmware of the 875ES and the 85032B manual. It's just the 4 fringing capacitance coefficients that seem different.

All other standards seem to match between the 85032B manual and the 8753ES firmware. But not the male open.

Sometime in the development of the kits I think we re-normalized the values between the 8753, 8720 and 8510. I know at one time the values for the 8753 were optimized for lower frequency (gave less ripple on a long-line test then the 8510 coefficients). At some later date the values were updated and I think they were made all the same. It might be an interesting experiment (for you) to compute the phase response of the two sets of coefficients and see how much they differ vs. frequency.