len vs key count frequency

#!/usr/bin/env python
#
# Efficiency of len against dict w/ key counting frequency
#
# Inspired in this answer http://stackoverflow.com/a/25919282/841339
#

from urllib import urlopen
from string import printable
from time import time

chars = {i: 0 for i in printable}
chars['\xef'] = 0
chars['\xbb'] = 0
chars['\xbf'] = 0

text = urlopen('http://www.gutenberg.org/cache/epub/4280/pg4280.txt').read()

dict_s = time()
for i in text:
    chars[i] += 1
sum(chars.values())
dict_e = time()

len_s = time()
len(text)
len_e = time()

print len_e-len_s, dict_e-dict_s