Checkpoint parameter works. When issuing a CHECK command in all sessions, their state looks identical.
It would actually be better to be consistent across all IMAP commands, but especially when NOOP or IDLE is used. Some servers have inconsistent IMAP session states because the client connections end up in, e.g., different servers that don't synchronize immediately. This can cause problems with clients that rely on the different sessions seeing the same state.
Exactly one session sees a new message as \Recent - no more and no less.
ImapTest prints an error if it notices multiple sessions having the same \Recent flag ("Message seq=.. UID=.. has \Recent flag in multiple sessions"). This happens automatically when running ImapTest with multiple clients for the same user, e.g., imaptest user=testuser clients=10.
But missing \Recent flags isn't detected with any tests.
Number of failures using scripted tests. These numbers may not be exact all the time, because the tests are still changing.
Failure groups: Each test belongs to a wider group of tests, typically testing a command or part of a command. If this count is low but individual command failure count is high, it probably means that the server has failed to implement wrong only a couple of commands.
Base fails: Number of individual base IMAP4rev1 protocol commands that failed.
Ext fails: Number of individual IMAP extension commands that failed. Extensions not supported aren't included in the numbers.
Test using, e.g., imaptest user=testuser test=tests/.