Benchmarks · /benchmarks/arena-video-edit

Video Edit Arena

Blind preference arena for instruction-based video editing quality.

Source · Arena
Version · arena snapshot 2026-05-13
Scores · 6

Passport

Visible tradeoffsThis is a human preference signal, so it tells you what people liked side by side, not what is formally correct.

source

Arena

metric

Arena rating (rating)

judge

Human

direction

higher better

group id

arena_video_edit_2026_q2

domain

Video generation

What it measures vs what it misses

✓ Measures

Observed user preference when models edit the same input video. Relative success at preserving context while applying requested changes.

✗ Misses

Shot-level edit faithfulness metrics. Audio continuity, render speed, and production-grade control surfaces.

Why this countsObserved user preference when models edit the same input video.Comparable-group ruleThis percentile only compares models inside the exact benchmark/version group shown here. It is not a universal score.What it missesShot-level edit faithfulness metrics.

Leaderboard · this benchmark version

#1 · dreamina-seedance-2.0-720p

AR · May 13, 2026

1,379

#2 · happyhorse-1.0

AR · May 13, 2026

1,319

#3 · Grok Imagine Video

AR · May 13, 2026

1,265

#4 · kling-o3-pro

AR · May 13, 2026

1,247

#5 · kling-o1-pro

AR · May 13, 2026

1,208

#6 · runway-gen4-aleph

AR · May 13, 2026

1,204