INF = 1 << 60

n, i = map(int, input().split())

dp = [- INF for _ in range(i + 1)]
dp[0] = 0
for _ in range(n):
  s, a = map(int, input().split())
  for j in range(i - s, -1, -1):
    if dp[j] >= 0:
      dp[j + s] = max(dp[j + s], dp[j] + a)

print(max(dp))