VEFX-Bench: leaderboard

VEFX-Bench (2026) compares instruction-guided video editing systems on curated source videos and editing prompts, scoring each edit with a reward model trained on human-annotated edits. This board is the overall GeoAgg score on a 1 to 4 scale, a per-edit weighted geometric aggregate of instruction following, rendering quality and edit exclusivity; higher is better.

Metric: Overall GeoAgg score (1-4; per-sample weighted geometric aggregate of instruction following, rendering quality and edit exclusivity with instruction following weighted twice, averaged over the evaluated items) on the 300 curated VEFX-Bench source-video and instruction pairs, each system's edited video scored by the authors' VEFX-Reward-32B evaluator from soft expected predictions; higher is better. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 10 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Kling 3.0 Omni3.06
2Kling o12.99
3Runway Gen-4.52.91
4Seedance 2.02.77
5Grok Imagine2.72
6Luma ray 32.72
7UniVideo2.52
8Wan2.62.15
9Luma ray 21.8
10VACE (VEFX-Bench checkpoint unspecified)1.78

Interactive version: theaggregate.ai/benchmark?slug=vefx-bench · How It Works · Data refreshed daily, snapshot 2026-11-22.