ChartArena - Box Plot (Chinese): leaderboard
Metric: mAP-high (%) on Chinese box plots, each model parses the chart image into its native structured format, normalized format-agnostically to (header, entity, value) triples for numeric charts or to labeled graphs and trees for flowcharts and mind maps; mean average precision over similarity thresholds at the high tolerance level, averaged over three visual styles (digital renderings, printed photos, hand-drawn photos); higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 43 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Seed 2.0 Pro (Non-reasoning) | 55.2 |
| 2 | Qwen 3.5 35B A3B (Thinking) | 50.6 |
| 3 | Kimi K2.5 (Non-reasoning) | 47.6 |
| 4 | Gemini 3.1 Pro (Preview) | 45.2 |
| 5 | GLM-4.5V | 37.4 |
| 6 | MiMo-V2-Omni | 30.3 |
| 7 | Qwen 3 VL 4B (Thinking) | 29 |
| 8 | Qwen 3 VL 8B (Thinking) | 25.9 |
| 9 | Gemini 2.5 Pro | 22.1 |
| 10 | Qwen 3 VL 4B Instruct | 20 |
| 11 | Qwen 3.5 9B | 18.1 |
| 12 | Qwen 2.5 VL 32B Instruct | 16.4 |
| 13 | Qwen 2.5 VL 72B Instruct | 15.3 |
| 14 | Qwen 3 VL 235B A22B Instruct | 14.1 |
| 15 | GPT-5 | 12.8 |
Interactive version: theaggregate.ai/benchmark?slug=chartarena-box-plot-chinese · How It Works · Data refreshed daily, snapshot 2026-09-29.