Failure types

topology

wrong connections, as in a molecule's bonds or a circuit

Seen in 3 results, across 2 models and 1 task, on 2026-09-29. More than one model, but only one task, so it may still be the task that is hard, and the models may be fine.

A coffee shrub hedge-pruned flat on one side, by Gemini 3.8 Flash
Gemini 3.8 FlashA coffee shrub hedge-pruned flat on one side · 2026-09-29 · 0 of 3 checksAlso completeness
A coffee shrub hedge-pruned flat on one side, by Gemini 3.8 Flash
Gemini 3.8 FlashA coffee shrub hedge-pruned flat on one side · 2026-09-29 · 0 of 3 checksAlso completeness
A coffee shrub hedge-pruned flat on one side, by DeepSeek 4.1 Flash
DeepSeek 4.1 FlashA coffee shrub hedge-pruned flat on one side · 2026-09-29 · 0 of 3 checksAlso completeness

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

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