CADGenBench — leaderboard
CADGenBench evaluates model capability on multimodal tasks from the linked upstream source with Aggregate CAD score as the primary reported metric.
Metric: Aggregate CAD Score. Source: huggingface.co. Status: saturation imminent. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol (xHigh) | 0.53 |
| 2 | GPT-5.5 (xHigh) | 0.46 |
| 3 | Claude Fable 5 | 0.45 |
| 4 | GPT-5.6 Sol | 0.41 |
| 5 | GPT-5.5 Pro | 0.39 |
| 6 | Claude Opus 4.7 | 0.37 |
| 7 | GPT-5.5 | 0.36 |
| 8 | Claude Opus 4.8 | 0.35 |
| 9 | Gemini 3.1 Pro (Preview) | 0.31 |
| 10 | Claude Opus 4.6 | 0.31 |
| 11 | Claude Sonnet 4.6 | 0.28 |
| 12 | Gemini 3.1 Flash Lite | 0.25 |
| 13 | GLM-4.6V | 0.18 |
| 14 | Qwen 3 VL 235B A22B Instruct | 0.17 |
Interactive version: theaggregate.ai/benchmark?slug=cadgenbench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.