I think my opinion of the 85mm Sonnar ZM was unjust. I called it 'overpriced' in June 2008, and said one might as well get the 90mm APO Summicron-M ASPH instead for ~20% more moolah. I've since reformed that position based on information that I hadn't digested at the time.
This week I was going over some of Erwin Puts' material on Zeiss lenses. Specifically this piece. It's not new, but I read it carefully this time. ;^)
It looks strange at first, but I think the 2/85 Sonnar is growing on me.
Erwin's comparison to the 90mm APO Summicron-M ASPH points out three difference, two of which I hadn't considered before. The first one is obviously the APO feature versus not, whereby the Sonnar is said to show slight blue fringes at defocused edges (not documented). I imagine the effect is much like what the 2/100 Makro-Planar ZF does at f/2.
The second difference is that the Sonnar incorporates a floating-element design, a fact which is confirmed by the Zeiss datasheet. This grants it much better performance at close range compared to the Summicron, even in the "waist/headshot" range as shown in the test samples. This level of performance is found in the 75mm APO Summicron-M ASPH also.
The third difference is that the Sonnar and Summicron simply render out-of-focus backgrounds completely differently. I personally prefer the Sonnar's rendition in Erwin's examples. I reproach it the "cat eye" effect, which the Summicron doesn't exhibit at all, but would still choose it over the Summicron's "sharp blur" almost every time.
What's my new opinion?
I think the APO Summicron could still be a better lens if money is no object to you, but I might personally lean for the Sonnar myself. The main reason is actually not mentioned above... I would prefer the Sonnar because it would match color and drawing style with other Zeiss T* lenses which (them) I strongly prefer over their Leica counterparts per the dollar.