ChartJudgeBench - CPA Aspect Ratio: leaderboard
Metric: Pairwise judging accuracy (%; the 182 aspect ratio pairs; two-pass position-swapped protocol: each of ChartJudgeBench's 1,003 human-filtered chart pairs (an original chart and a copy with one injected flaw) is judged twice with the order reversed and counts as correct only if the judge picks the better chart both times; temperature 0.1). Source: arxiv.org. Saturation forecast: Around December 2026. 26 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro | 89 |
| 2 | Qwen 3 VL 32B Instruct | 69.8 |
| 3 | GPT-5.2 | 66.5 |
| 4 | Claude Sonnet 4.5 | 59.3 |
| 5 | Qwen 2.5 VL 72B Instruct | 54.9 |
| 6 | Qwen 3 VL 8B Instruct | 35.2 |
| 7 | Qwen 3 VL 30B A3B (Thinking) | 18.7 |
| 8 | Qwen 2 VL 7B Instruct | 11.5 |
| 9 | Qwen 2.5 VL 7B Instruct | 1.7 |
Interactive version: theaggregate.ai/benchmark?slug=chartjudgebench-cpa-aspect-ratio · How It Works · Data refreshed daily, snapshot 2026-09-26.