Note: Lens adapters have been shown to exacerbate or even cause issues along the edges and corners of some lenses due to an imperfect alignment with the image sensor. Roger Cicala reported on this in a blog post some time ago. It's worth mentioning here, because the Sigma is adapted to the Sony a7R II via a Metabones smart lens adapter in order to compare it directly to the Sony 85mm GM.  

As with DxO's measurements, in our real world tests we see that the Sigma 85mm F1.4 is extremely sharp centrally, when shot wide-open at F1.4, but the Sony 85mm GM isn't too far behind. In the corners, the Sigma is a bit sharper than the Sony GM when shot wide open, which also cocorroborates the DxO data. The Sigma is sharpest when stopped down to F4 centrally and at F5.6 in the corners, which reflects what we saw in the test data as well. The Sony sharpens up nicely as well by F4, but still falls behind the Sigma in terms of overall sharpness.

The Sony is also a bit softer in the corners than the Sigma across all apertures and doesn't sharpen up quite as nicely, as you stop the lens down, as also confirmed by the DxO data. Diffraction starts to come into play by F8 in both lenses, with more dramatic fall-off in terms of sharpness occurring after F11.

The Sigma handles lateral CA a bit better wide-open when compared to the Sony 85mm, which is also reflected in the DxO data. The CA is completely eliminated in both lenses by F4, with a marked decrease by F2.8. Longitudinal CA is handled extremely well by both of these lenses in this widget, with purple fringes only really visible wide-open in high contrast regions

In terms of vignetting; DxO reports a 1.9 stop falloff in the Sony and our real world example does see more vignetting when compared to the Sigma at F1.4. This vignetting improves as the lens is stopped down to F4, as the DxO data suggests, but the Sigma still performs a bit better than the Sony. This is a purely qualitative observation, but it does seem to correlate with the DxO data.