Tamron introduces 24-70mm f/2.8 with IS

[Mt Spokane Photography]

The multiple stage AF might be a result of this, the camera might do extra accurate AF for a 1.2 lens.

That sounds plausible, but would be bad because the f2.8 lens could gain speed with another lens code. However, a video review on youtube shows the same effect with a Nikon body, but maybe for the same reason. It's hard to tell without a thorough test.

I doubt that the lens would identify itself as a Canon 85mm f/1.2 to a Nikon body, Nikon codes are arranged a bit differently, so its hard to say what might be going on with Nikon bodies.
http://www.exiv2.org/tags-nikon.html

 

[Marsu42]

I doubt that the lens would identify itself as a Canon 85mm f/1.2 to a Nikon body

No, I wasn't implying that - only that it identified as an "larger than f2.8" aperture Nikon lens, so the af might show the same behavior as on Canon bodies. But since we cannot tell the lens to change its id, this doesn't matter anyway - the Tamron af is as it is, it's a little slower than Canon but still very capable in a sub-€1000 package.

 

[Marsu42]

I wonder if the code it uses has been selected such that the camera will turn on all of the 2.8 dual cross type auto focus sensors?

This was just discussed in another thread: http://www.canonrumors.com/forum/index.php?topic=7246

In short: Yes, the 137 lens code should activate all double-cross af sensors (with the possible loss of af speed), but the question is if there was a reason besides marketing for Canon to only enable the central one for the 24-70mk1 like loss of sharpness in mid-frame which shouldn't occur on the expensive 24-70mk2.