Recognises undecided history requirements across the register
Row 10 across the register: for each significant thing, nobody has answered whether a past report must be reproducible exactly as published.
Can AI do this?
People still lead Few or none of the tested AI models showed they master this. 1 of 10 models master it.
| Model | Tier | Probability of mastery | Status |
|---|---|---|---|
| gemini-4-argon | Flagship | 91% | Mastered |
| gemini-3.1-pro-preview | Flagship | 32% | Not shown |
| gpt-6-astra | Flagship | 81% | Not shown |
| claude-fable-5-1 | Flagship | 82% | Not shown |
| gemini-3.8-flash | Mid tier | 35% | Not shown |
| gpt-6.1-sol | Mid tier | 80% | Not shown |
| claude-sonnet-5-5 | Mid tier | 81% | Not shown |
| gemini-3.5-flash-lite | Low cost | 31% | Not shown |
| gpt-6-luna | Low cost | 36% | Not shown |
| claude-haiku-5-5 | Low cost | 34% | Not shown |
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: The significant things whose history need is undecided
How it is assessed: Flags the deficient case and passes the acceptable one. Deficient: Rate, status and address requirements with no statement either way Acceptable: One answer per thing, recorded as a requirement or as an open_question that blocks the logical model.
Also known as: History need not settled (data requirement)
Related knobits
Builds on
- K-110 Recognises a data requirement silent on current versus past values
- K-155 Explains current value, history and point-in-time reproducibility
All 47 knobits of Capture data requirements.
Source: Review checklist 01-requirements, row 31; Stage 1 (time)