Failure types

geographic-grounding

coordinates that don't match real streets or places

Seen in 1 result, across 1 model and 1 task, on 2026-09-20. One model so far, which is not enough to call it a pattern.

A run that spells COFFEE, by GLM 5.3 FlashX
GLM 5.3 FlashXA run that spells COFFEE · 2026-09-20 · 0 of 5 checksAlso physical-plausibilityDrawn north-up across the blocks and out of the grid entirely, and still not readable as COFFEE.

The other types: labelling · physical-plausibility · completeness · part-order · topology

All of them, with counts · How a check is decided