PaintBench - Geometric Transformation: leaderboard

PaintBench mIoU (%) averaged over its five geometric transformation tasks (translation, rotation, reflection, scaling and shearing); edits are scored pixel by pixel against the single correct image, higher is better (NYU, 2026).

Metric: mIoU (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 11 models tracked.

Top models

#ModelScoreOverall rank
1GPT-Image-211.1
2Nano Banana6.2
3Nano-Banana-26.1
4Qwen-Image-Edit-25113.4
5FLUX.2-dev3.3
6BAGEL-7B-MoT2.4#5515
7FLUX.1-Kontext-dev2.4
8LongCat-Image-Edit2.2
9FLUX.2-klein-9B1.4
10HunyuanImage-3.00.1

Interactive version: theaggregate.ai/benchmark?slug=paintbench-geometric-transformation · How It Works · Data refreshed daily, snapshot 2026-11-12.