The short version
A model benchmark that rewards reliable knowledge and penalizes confident hallucinations.
From the HackerLinks archive
Artificial Analysis Omniscience Index
A model benchmark that rewards reliable knowledge and penalizes confident hallucinations.
At a glance:
First seen:2026-09-06
Last seen:2026-09-06
Times seen:1
Website:artificialanalysis.ai
Why it caught our attention
A commenter said it tracks practical model usefulness better than broad benchmark scores.
Where it surfaced on Hacker News
2026-09-06
“Imo the omniscience index they have has the highest correlation to actual usefulness of the models.”
jascha_eng · recommendation · evaluated
Direct HN comment
Artificial Analysis Intelligence Index v4.2
Also surfaced in this discussion
CursorBench
Artificial Analysis Intelligence Index v4.2
2026-09-06