ChartSync: leaderboard
ChartSync (2026) tests image editing models on 870 statistical-chart edits in which changing a value must also update the geometry that depends on it. The overall score averages OCR text fidelity, SSIM and three VLM-judge scores, in percent; higher is better.
Metric: Overall score (%; equal average of five metric-level scores: character-level OCR F1 of all chart text, SSIM against the ground truth, and three Gemini-3.1-Pro judge scores for textual edit success, visuo-logical consistency and background fidelity; 870 chart editing triplets (original chart, instruction, programmatically rendered ground truth) over 9 chart categories, 235 of them geometry-coupled cascading edits; each model edits the chart image once). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Nano Banana Pro (Gemini 3 Pro Image) | 87.76 | |
| 2 | GPT-Image-2 | 81.78 | |
| 3 | Wan2.7-Image-Pro | 74.47 | |
| 4 | SeeDream-5.0-Lite | 71.09 | |
| 5 | Qwen-Image-2.0-Pro | 65.54 | |
| 6 | Qwen-Image-Edit-2511 | 60.29 | |
| 7 | FireRed-Image-Edit-1.1 | 56.22 | |
| 8 | FireRed-Image-Edit-1.0 | 51.61 | |
| 9 | Qwen-Image-Edit | 50.75 | |
| 10 | FLUX.2-klein-base-9B | 50.62 |
Interactive version: theaggregate.ai/benchmark?slug=chartsync · How It Works · Data refreshed daily, snapshot 2026-11-13.