0

LMArena - Instruction Following

Human-preference Elo on prompts that carry explicit constraints - does the model actually do what the brief asked.

Open
Operator
LMArena
Kind
Human preference
Updates
weekly·updated 15h ago
Notable for
human-preference
URL
arena.ai/leaderboard/text/instruction-following
Tracks
349 models · Elo

Cite

Notes

Only stored in your browser.

Attribution

Attribution policy →

Elo ranking

349

models

LMArena - Instruction Following · EloBar chart with 21 bars. Highest value: Claude Opus 4.6 at 1523.
21 models

Backing benchmark

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