I decided to go ahead and compare it using a loop of 10,000 quads.
Results:
Method 1: 0.0128 seconds
Method 2: 0.0132 seconds
Method No. 1 has some improvement, but the improvement is very slight (3%). This is probably nothing more than overhead, just calling more features. Thus, it is likely that OpenGL itself does not receive additional optimization from method # 1.
3 Windows XP OpenGL 2.0 visual studio 2005.