GAHHH. They messed up the units! The stated numbers (4 MPa for imperfect graphene) are wrong! They are meant to be fracture toughness (which is measured in Pa/m^0.5), not just Pa. The paper is at [0]

Table 1 is where you want to look at. It places the fracture stress of pre-cracked graphene in the 1-10 GPa range(which should be compard to the stated 100 GPa for perfect graphene).

I don't have the full materials properties of graphene with me, and I'm not a materials guy, so I may have done something wrong above (so please correct me), but I'm fairly certain that at a minimum, the article messed up the units.

[0] http://www-old.me.gatech.edu/zhugroup/paper/NC14.pdf

EDIT: replaced the link to a copy of the paper hosted by the author from the original nature (paywalled) link.

EDIT2: some changes after actually reading the paper instead of going off the abstract and supplied figures.