IBM Assembler supports binary (fixed-point), decimal (packed or zoned), and floating-point arithmetic instructions to handle different types of numeric data. Since these three options cover the valid arithmetic modes, 'None of the Above' is incorrect.