Answer: 4 bytes (32 bits).
A float stores single-precision real numbers. On practically every C compiler (Turbo C, GCC, Visual C++) it takes 4 bytes, following the IEEE 754 single-precision format: 1 sign bit, 8 exponent bits and 23 bits for the fraction.
This gives about 6–7 significant decimal digits and a range of roughly \(1.2\times10^{-38}\) to \(3.4\times10^{38}\) (in magnitude). For more precision use double (8 bytes). You can confirm with sizeof(float).