Yes, it was a mistake. We had anti tests for the brute forces. But we had only one such test in a file which surely was not enough to stop them (we thought it was enough).
It was also partly due to me asking to try to reduce the test files and constraints a bit to reduce load on servers. We reduced constraints on sum of lengths after everything was prepared.We messed up somewhere around there.
Author is updating tests for practice version of the problem.