CADGenBench: leaderboard
Validated CADGenBench v0.1.0 submission aggregate CAD scores, preserving submission, benchmark revision and agent configuration. Generation and editing include validity-gated geometric, interface and topology criteria.
Metric: Aggregate CAD Score. Source: github.com. Source: huggingface.co. Saturation forecast: Around July 2028. 25 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 (xHigh) | 0.46 | #31 (GPT-5.5) |
| 2 | Claude Fable 5 | 0.45 | #16 |
| 3 | GPT-5.6 Sol | 0.41 | #19 |
| 4 | GPT-5.5 Pro | 0.39 | #17 |
| 5 | Claude Opus 4.7 | 0.37 | #49 |
| 6 | GPT-5.5 | 0.36 | #31 |
| 7 | Claude Opus 4.8 | 0.35 | #37 |
| 8 | Gemini 3.1 Pro (Preview) | 0.31 | #58 |
| 9 | Claude Opus 4.6 | 0.31 | #63 |
| 10 | Claude Sonnet 4.6 | 0.28 | #90 |
| 11 | Gemini 3.1 Flash Lite | 0.25 | #191 |
| 12 | GLM-4.6V | 0.18 | #316 |
| 13 | Qwen 3 VL 235B A22B Instruct | 0.17 | #273 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=cadgenbench · How It Works · Data refreshed daily, snapshot 2026-10-20.