I really appreciate your detailed response.
I tried changing the key to a column with higher cardinality—a column that includes both date and time (timestamp)—to achieve a more even distribution.
Now the size has dropped to 4.3 GiB, and the number of files is 64.

Currently, the queries performed similarly, with a slight advantage for the table with the CLUSTER BY, when filtering on the clustered column. However, the amount of MB read was still higher.
I was under the impression that, unlike PARTITION BY, I don’t need to find a column with an even distribution for CLUSTER BY.

Could you clarify in which cases I would see a significant improvement with a CLUSTER BY table compared to a table with no optimization strategy at all?

Thank you!