// #pragma GCC target("avx2")
// #pragma GCC optimize("O3")
// #pragma GCC optimize("unroll-loops")
#include <bits/stdc++.h>
// #include <atcoder/all>


using namespace std;
// using namespace atcoder;
#define rep(i,m,n,k) for (int i = (int)(m); i < (int)(n); i += (int)(k))
#define rrep(i,m,n,k) for (int i = (int)(m); i > (int)(n); i += (int)(k))
#define all(x) x.begin(),x.end()
#define ll long long
#define list(T,A,N) vector<T> A(N);for(int i=0;i<(int)(N);i++){cin >> A[i];}
template<typename T> inline bool chmin(T& a, T b){if(a > b){a = b; return true;} return false;}
template<typename T> inline bool chmax(T& a, T b){if(a < b){a = b; return true;} return false;}
// #define int long long


int main(){
    ios::sync_with_stdio(false);
    std::cin.tie(nullptr);
    int N,V,sx,sy,gx,gy;
    cin >> N >> V >> sx >> sy >> gx >> gy;
    vector<vector<int>> L(N,vector<int> (N));
    rep(i,0,N,1){
        rep(j,0,N,1){
            cin >> L[j][i];
        }
    }
    vector<pair<int,int>> dxy = {{-1,0},{1,0},{0,1},{0,-1}};
    vector<int> val(200000000);
    sx -=1;
    sy -=1;
    gx -=1;
    gy -=1;
    int N2 = N*N;
    val[sx+N*sy+N2*V] = 1;
    queue<int> v;
    v.push(sx+N*sy+N2*V);
    while(!v.empty()){
        auto xym = v.front();
        v.pop();
        auto m = xym/N2;
        auto y = xym%N2/N;
        auto x = xym%N;
        for(auto [dx,dy]:dxy){
            if((x+dx>=N)||(y+dy>=N)||(x+dx<0)||(y+dy<0)) continue;
            if(m-L[x+dx][y+dy]<=0) continue;
            if(val[x+dx+N*(y+dy)+(m-L[x+dx][y+dy])*N2]!=0) continue;
            if((x+dx==gx)&&(y+dy==gy)){
                cout << val[xym] << endl;
                return 0;
            }
            val[x+dx+N*(y+dy)+(m-L[x+dx][y+dy])*N2] = val[xym]+1;
            v.push(x+dx+N*(y+dy)+(m-L[x+dx][y+dy])*N2);
        }
    }
    cout << -1 << endl;

    return 0;

}