ChartArena - Mind Map (English): leaderboard
Metric: mAP-high (%) on English mind maps, 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 December 2026. 40 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 86.8 |
| 2 | Seed 2.0 Pro (Non-reasoning) | 83.1 |
| 3 | Kimi K2.5 (Non-reasoning) | 80.8 |
| 4 | GPT-5 | 76.6 |
| 5 | MiMo-V2-Omni | 76.6 |
| 6 | Gemma 4 31B (IT) | 76.2 |
| 7 | Qwen 3.5 35B A3B (Thinking) | 75.1 |
| 8 | Gemini 2.5 Pro | 71.7 |
| 9 | Qwen 3 VL 235B A22B Instruct | 70.8 |
| 10 | Qwen 3 VL 8B Instruct | 66.4 |
| 11 | GLM-4.5V | 66.2 |
| 12 | Qwen 3 VL 8B (Thinking) | 64.8 |
| 13 | Qwen 3.5 9B | 64.2 |
| 14 | GPT-4o | 64 |
| 15 | Qwen 2.5 VL 72B Instruct | 63.8 |
Interactive version: theaggregate.ai/benchmark?slug=chartarena-mind-map-english · How It Works · Data refreshed daily, snapshot 2026-09-29.