float has less precision than double, which would be the default type used for floating point literals. Since 6.7 cannot be represented with a finite number of binary digits, the less precise float representation does not equal the double representation.
solved Testing for float equality in C [duplicate]