// #pragma GCC optimize("O3,unroll-loops")
#include <bits/stdc++.h>
// #include <x86intrin.h>
using namespace std;
#if __cplusplus >= 202002L
using namespace numbers;
#endif
template<class T> T &ctmin(T &x){ return x; }
template<class T, class Head, class ...Tail> T &ctmin(T &x, const Head &h, const Tail &... t){ return ctmin(x = min<T>(x, h), t...); }
template<class T> T &ctmax(T &x){ return x; }
template<class T, class Head, class ...Tail> T &ctmax(T &x, const Head &h, const Tail &... t){ return ctmax(x = max<T>(x, h), t...); }



int main(){
	cin.tie(0)->sync_with_stdio(0);
	cin.exceptions(ios::badbit | ios::failbit);
	static const vector<pair<int, int>> dr8{{2, 1}, {1, 2}, {1, -2}, {-2, 1}, {-2, -1}, {-1, -2}, {1, -2}, {2, -1}};
	int n;
	cin >> n;
	set<array<int, 2>> appear;
	map<array<int, 2>, int> mask;
	for(auto i = 0; i < n; ++ i){
		int x, y;
		cin >> x >> y;
		appear.insert({x, y});
		for(auto [dx, dy]: dr8){
			mask[{x + dx, y + dy}] |= 1 << i;
		}
	}
	for(auto [x, y]: appear){
		mask.erase({x, y});
	}
	const int inf = numeric_limits<int>::max() / 2;
	vector<int> dp(1 << n, inf);
	dp[0] = 0;
	for(auto [pos, m]: mask){
		for(auto smask = m; smask; smask = smask - 1 & m){
			ctmin(dp[smask], 1);
		}
	}
	for(auto mask = 1; mask < 1 << n; ++ mask){
		for(auto smask = mask; smask; smask = smask - 1 & mask){
			ctmin(dp[mask], dp[smask] + dp[smask ^ mask]);
		}
	}
	dp.back() == inf ? cout << "-1\n" : cout << dp.back() << "\n";
	return 0;
}

/*

*/