1print (grouped.groupby(level='yearmonth').nlargest(3).reset_index(level=0, drop=True))
2yearmonth product
3201601 E 180
4 A 100
5 B 90
6201602 F 220
7 A 200
8 C 120
9201603 B 120
10 C 110
11 D 110
12Name: val, dtype: int64