See GR-MEGA manual page 42 - 46, in a section called “Granular Slice Sound Engine”.
I assume you are aware of the GR-2 ( thread ) that is close to half the price of the MEGA. You mention cost, size, and not caring about the sequencer, all in the GR-2’s favor.
To clear up the confusion about slicing on the GR-MEGA: yes, it could already do that since 1.0. Since 1.7 it can even load SFZ files (multisamples)… Only it doesn’t support velocity layers and such. Full multisample support is coming in 1.10 as you already posted here. Hope this helps! Thanks for your support!