In both Python 2 and 3, unambiguously define a sentinel Unicode string (&Ж中𐍆).

unitest.py

# encoding: utf8
"""
Unambiguously define, from hex data, a sentinel Unicode string (&Ж中𐍆).

The string contains a 1-, 2-, 3-and 4-byte Unicode character. Useful for finding
encoding errors in editors and transfer systems!
"""
import sys
py2 = True if sys.version_info.major == 2 else False

codes = "26  d096  e4b8ad  f0908d86"
bytes_ = [codes[i:i+2] for i in range(0, len(codes), 2)]
bytes_ = [int(byte, 16) for byte in bytes_ if byte.strip()]
if py2:
    bytes_ = [chr(char) for char in bytes_]
    unistr = unicode(''.join(bytes_), 'utf8', 'strict')
    unistr = unistr.encode('utf8')
    print unistr    
else:
    unistr = bytes(bytes_).decode('utf8')
print(type(unistr), unistr)