0

Arena Vision

LMArena subcategory for vision-language models (VLMs), ranked by user pairwise votes on prompts that include an image.

Operator
LMArena
Kind
Human preference
Updates
weekly·updated 5d ago
Notable for
The public preference benchmark for vision-language models, complementing structured VLM benchmarks like MMMU and MathVista.
Tracks
112 models · Elo

Cite

Notes

Only stored in your browser.

Elo ranking

112

models

Arena Vision · EloBar chart with 21 bars. Highest value: Claude Opus 4.7 at 1349.
21 models

Backing benchmark

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