The evidence for single-gender vs. coeducational schools is mixed. The trend over the past few centuries is towards full integration of the classroom. There are several reasons behind this.
Philosophically, most educators believe that students benefit from diversity. Having students from both genders and a wide variety of ethnic, religious, and socioeconomic backgrounds in the same classroom benefits the students by making them comfortable with a wide range of peoples, preparing them for life in a mobile and diverse world and building up a wide range of sympathies and friendships with people quite different from themselves. There is some evidence that the more closely boys and girls work together on school projects, the less likely they are to believe in sexual stereotypes.
Another issue is that in some countries, if schools are segregated by gender, girls will be shortchanged, given fewer resources and less funding across the board than boys.
On the other side of the equation, girls at women's colleges in the United States are more likely to graduate on time, go on to graduate and professional schools, and attain leadership positions in business and politics than women educated at coeducational institutions.
In the long run, I do not think that either of these two choices is uniformly better, but rather that some children respond better to single-gender education and some to coeducation, and that different educational systems work better in different cultural milieu. For example, in countries such as Afghanistan and Pakistan, girls' schools provide safe havens for young women who might otherwise be denied an education, but coeducation is quite successful in the more egalitarian Nordic countries.