UniEditBench (Image) - Text Alignment: leaderboard
UniEditBench (2026) scores image edits on 633 samples across nine editing operations on a 1-5 scale with an evaluator the authors distilled from Qwen3-VL-235B-A22B into Qwen3-VL-8B. This board rates how well the edit follows the instruction, from 1 to 5; higher is better.
Metric: Text alignment score (1-5): the edit does what the instruction asks; 633 image-editing samples over nine operations, each edit scored on a 1-5 Likert scale by the authors' Qwen3-VL-8B evaluator distilled from Qwen3-VL-235B-A22B (the second entry of each 4B/8B cell); higher is better. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 16 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | UniEditBench Hunyuan (checkpoint unspecified) | 4.61 | |
| 2 | Qwen-Image-Edit | 4.35 | |
| 3 | Step1X-Edit | 3.87 | |
| 4 | UniWorld-V2 | 3.85 | |
| 5 | FLUX.1-Kontext-dev | 3.71 | |
| 6 | BAGEL-7B-MoT | 3.7 | #4668 |
| 7 | OmniGen2 | 3.6 | #4625 |
| 8 | LaVida-O | 3.59 | |
| 9 | VAREdit | 3.55 | |
| 10 | DreamOmni2 | 3.48 | |
| 11 | ICEdit | 3.01 | |
| 12 | EditMGT | 2.75 | |
| 13 | EditAR | 2.55 | |
| 14 | AnyEdit | 2.43 | |
| 15 | control_v11e_sd15_ip2p | 1.8 |
Interactive version: theaggregate.ai/benchmark?slug=unieditbench-image-text-alignment · How It Works · Data refreshed daily, snapshot 2026-11-22.