GRADE (Image Editing) - Relaxed Score: leaderboard

Metric: Relaxed score (0-100): 0.6 x discipline reasoning + 0.3 x visual consistency + 0.1 x logical readability, each normalized to 0-100, on GRADE's 520 discipline-informed image editing samples (input image, editing instruction and ground-truth image across 10 academic disciplines), a Gemini-3-Flash judge scoring reasoning against expert-verified weighted questions (written by GPT-5) and visual consistency and logical readability on 0/1/2 scales; higher is better. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 20 models tracked.

Top models

#ModelScoreOverall rank
1Nano Banana Pro82.9
2Nano Banana 279.1
3Seedream 5.0 Lite73.7
4GPT-Image-1.566.5
5Nano Banana56
6FLUX.2-max55.7
7GPT-Image-154.2
8Seedream 4.549.7
9FLUX 2 Pro47
10DreamOmni44.3

Interactive version: theaggregate.ai/benchmark?slug=grade-image-editing-relaxed-score · How It Works · Data refreshed daily, snapshot 2026-10-11.