TinyGrafixBench: leaderboard
Tiny graphical editing benchmark for small, precise changes to charts, diagrams, icons, and other structured visual artifacts.
Metric: Mean IoU (%). Source: paintbench.github.io. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Nano Banana 2 | 15.9 | |
| 2 | GPT Image Edit 2 | 15.6 | |
| 3 | Nano Banana 1 | 5.3 | |
| 4 | Qwen-Image-Edit | 4.3 | |
| 5 | FLUX.2-Klein-9B | 3.4 | |
| 6 | LongCat-Edit | 3.2 | |
| 7 | FLUX.2-dev | 3.1 | |
| 8 | FLUX.1-Kontext | 3.1 | |
| 9 | BAGEL-7B-MoT | 2.7 | |
| 10 | HunyuanImage-3 | 0.3 |
Interactive version: theaggregate.ai/benchmark?slug=tinygrafixbench · How It Works · Data refreshed daily, snapshot 2026-10-20.