IN / OR and by and large the same operation, for example:

SQL> set autotrace traceonly explain
SQL> select *
  2  from t
  3  where object_id in (123,456,789);

Execution Plan
----------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation         | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |      |     3 |   345 |   290   (1)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T    |     3 |   345 |   290   (1)| 00:00:01 |
--------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   1 - filter("OBJECT_ID"=123 OR "OBJECT_ID"=456 OR "OBJECT_ID"=789)

Notice the filter - the IN is mapped to OR anyway in this case.

The answer to your specific question is "it depends"

select * from test.testing d where D.code in ('123','124','136','136');

Out of the "millions of rows", how many of them have code 123,124, etc...?

If it's just a few rows, then an index on code would be a sensible option. If it's hundreds of thousands of rows, using an index on code would be a disasterous option.

Is it query that is run 100's of times per day ? Then we might to look at structures to make it more efficient ? Or it is run once per month - then maybe we dont care if it takes a while. Or is it run out on a server with lots of I/O bandwidth - then we might look at a parallel options for the SQL.