Explains unmerging master records
Reversing a merge: restoring both identities, recomputing both golden records and correcting every downstream consumer that received the merged identity.
Can AI do this?
AI does this well Most of the tested AI models master this. 4 of 4 models in the current benchmark master it.
| Model | Tier | Probability of mastery | Status |
|---|---|---|---|
| gemini-4-argon | Flagship | 99% | Mastered |
| gemini-3.5-flash-lite | Low cost | 98% | Mastered |
| gpt-6-luna | Low cost | 99% | Mastered |
| claude-haiku-5-5 | Low cost | 98% | Mastered |
Mastered means a probability of at least 90% that the model has this knobit, estimated from its graded answers. “Not shown” means the answers do not settle it either way. How the models are tested.
What it looks like in practice
Expected result: Reversing a merge: restoring both identities, recomputing both golden records and correcting every downstream consumer that received the merged identity.
How it is assessed: Answers a written question on the unmerge correctly
Related knobits
Needed for
- R-404 Recognises a golden record with no unmerge path
- R-405 Recognises an unmerge leaving downstream systems merged
All 14 knobits of Build golden records by survivorship.
Source: Practice the stage 4 reference and review checklist rest on