I've been looking at deaths I'd previously found on FreeBMD with ages '0' or '1'.
On the new GRO index the FreeBMD '0's are coming up as '0' or a number between 1 and 11.
FreeBMD '1's are coming up on the new GRO index as a mix of '1' or a number between 12 and 22 (not seen a 23 yet).
My hunch is the numbers where a '0' or '1' is expected are generally in months, but there may not be consistency.
I.e.
Up to 1 month = 0
Up to 1 year = a number representing between 1 and 11 months
Over 1 year but less than 2 years = either '1' or a number representing between 12 and 23 months
I'd be interested to hear of any known ages less than 1 month which are shown as a number of weeks/days.
This does appear to be a bit of a mess. Unfortuantely I fear the GRO will 'solve' the problem by editing the ages to '0' rather than changing the website to correctly search and display ages in days/weeks/months.

So I'd say it should be a priority to look at children who you know died at a very early age (i.e. you have already found them on FreeBMD with a '0' or '1') and make a note of the number in the age column in the new index. Working out whether it is days/weeks/months can happen later

That is what's kept me busy this weekend.