Currently this is hardcoded - and currently I don't see a situation where making this configurable would help:
* Greylisting is meant to catch spam, that is sent by a one-shot script instead of being queued on a SMTP server for delivery - these scripts try to send the mail only once, and don't bother retrying on seeing a 4xx temp-fail code
* Greylisting will not help to catch spam, which somehow was queued on a smtp-server (if for example the spammers are using postfix/exim/exchange for sending spam)
* Do you have an example of a spam-mail, which got retried after a few minutes, but would not have been retried after say 1h?