BANTUNOMICS L26 · Operating-Alphabet Test Leaderboard White paper Why FSI

The L26 Operating-Alphabet Test

Every model has seen the alphabet a trillion times. Ask it to recite the closed set — the operating alphabet of a language it claims to speak — and it stumbles. L26 scores that out of 26, the size of the English alphabet, so any language reads at a glance: 26/26 is mastery. Run your own model below.

1 · Choose the alphabet

Blind — distil it unaided Scaffolded — onsets given
Blind is the real test. Scaffolded hands over the inventory.
Paste output
Bring your API key

2 · Run this prompt in your model

Copy it into your model, then paste what it returns below. Nothing is shown to the model but the task.
…
L26 v1.0 · scored deterministically (no LLM judge) · the ground-truth inventory is never shown to the model (zero-contamination). Full method: the white paper. See where every lab stands on the public leaderboard.