Benchmarks · reasoning
Chartography
Can a model read the charts professionals rely on — Kaplan-Meier curves, candlesticks, contour maps, Sankey diagrams, Bode plots, wind roses — from the image alone, to an expert-set precision? Maintained by Surge AI. Score: The board's accuracy: the share of answers inside the expert-set acceptable range of the golden answer, judged per answer by Gemini 3.5 Flash, averaged over the 100 tasks and ten runs per task.
41 models · 61 results · data as of 2026-10-29 · surgehq.ai ↗
Leaderboard
GPT-6 Astra max
All results
61 rows| # | Model | Effort | Chartography |
|---|---|---|---|
| 1 | GPT-6 Astra | max | 71.0% |
| 2 | GPT-6 Astra | medium | 70.0% |
| 3 | Opus 5.5 | max | 66.3% |
| 4 | GPT-6 Sol | max | 53.6% |
| 5 | GPT-6 Sol | medium | 46.5% |
| 6 | Fable 5.1 | max | 46.2% |
| 7 | GPT-5.6 Sol | max | 45.0% |
| 8 | Gemini 3.7 Flash | medium | 43.0% |
| 9 | Gemini 3.8 Flash | medium | 42.5% |
| 10 | Gemini 3.8 Flash | high | 40.9% |
| 11 | Gemini 3.7 Flash | high | 40.4% |
| 12 | GPT-5.6 Sol | medium | 39.5% |
| 13 | Gemini 3.5 Flash | medium | 35.9% |
| 14 | Fable 5 | max | 34.8% |
| 15 | Gemini 3.6 Flash | medium | 34.0% |
| 16 | GPT-5.6 Terra | max | 34.0% |
| 17 | Muse Spark 1.2 | xhigh | 32.1% |
| 18 | Muse Spark 1.2 | max | 31.5% |
| 19 | GPT-5.6 Luna | max | 31.4% |
| 20 | GPT-5.5 | xhigh | 31.0% |
| 21 | GPT-5.4 | xhigh | 29.6% |
| 22 | GPT-6 Luna | max | 29.1% |
| 23 | Qwen 3.8 Max | xhigh | 29.1% |
| 24 | GPT-5.5 | medium | 28.5% |
| 25 | Muse Spark 1.3 | xhigh | 27.6% |
| 26 | Opus 5 | max | 27.3% |
| 27 | GPT-6 Luna | medium | 26.7% |
| 28 | GPT-5.6 Terra | medium | 26.6% |
| 29 | Kimi K3 | max | 26.6% |
| 30 | Gemini 3.1 Pro | high | 26.1% |