This would be very cool. Considering the wind, the wrong turn early in the race, the fact that Tokyo is not so fast as Berlin, and the lack of perfect long term pacing, Eliud's recent 2:02:40 in Tokyo does seem to indicate near WR shape. I think the wrong turn alone cost about 10 seconds; Tokyo may be at least 15-30 seconds slower than Berlin (possibly much more), and assuming the headwind was more to blame for Eliud's slowdown in the final 7k than just fatigue, we can easily imagine that performance was worth sub2:02 in Berlin, if not close to the record.
Perhaps they will have him open in 60:30-40, just so the possibility of smashing the record yet again is there, but even with fading in the latter stages he may improve it some. It's tough to open in 61:00 and count on a sub 60:39 second half to get there, although Eliud may be comfortable with such an ask.