Again I don't think this should be used for tiering at all. It's very vague, and you could probably interpret it in a way that wouldn't require scaling (as other Keter Duty type containment have managed to work without needing direct scaling).
And like I said, anything can be High 1-A if inserted in conceptual space. Such possibility isn't worth tiering at all unless this state is notable (Like SCP-2317 having actual, explicit feats of fighting off SCP-5800 as a whole). And 3812's case is way too vague to be worth a key. If anything, the statement we get indicates how small and insignificant 3812 is compared to 5800, making any scaling at all dubious at best.
On top of that, 3812 scaling to 5800 in such a way... doesn't make sense. Narrative layers that 3812 ascends would either be entirely below or entirely above 5800. Not in this weird state where 3812 can meaningfully fight with 5800 without instantly destroying them. So it's not only vague, but it's also mostly nonsensical. It really should be discarded unless we get more and better scaling material.