VEFX-Bench - Rendering Quality: leaderboard
Metric: Rendering-quality score as a coverage-adjusted model-level estimate (inverse-propensity-weighted mixed-effects) on the 1-4 rubric scale over 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.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Kling 3.0 Omni | 3.59 |
| 2 | Kling o1 | 3.53 |
| 3 | Seedance 2.0 | 3.42 |
| 4 | Luma ray 3 | 3.4 |
| 5 | Grok Imagine | 3.35 |
| 6 | Runway Gen-4.5 | 3.32 |
| 7 | Wan 2.6 | 3.32 |
| 8 | UniVideo | 3.27 |
| 9 | VACE (VEFX-Bench checkpoint unspecified) | 3.17 |
| 10 | Luma ray 2 | 2.53 |
Interactive version: theaggregate.ai/benchmark?slug=vefx-bench-rendering-quality · How It Works · Data refreshed daily, snapshot 2026-10-07.