0

Arena Hard Prompts

LMArena subcategory ranking models on a filtered slice of Arena prompts auto-classified as hard along multiple difficulty axes.

Operator
LMArena
Kind
Human preference
Updates
live
Notable for
The user-preference view of frontier model separation on hard prompts; correlated with Arena-Hard offline benchmark.
Tracks
Preference voting (no benchmark)

Cite

Notes

Only stored in your browser.

Backing benchmark

Human-preference voting. No underlying benchmark - models are ranked by pairwise votes, not by a test you can run.