CadQueryEval Corrected Geometry Hausdorff
Code · 2026-09-30
CadQueryEval final corrected-geometry release: percentage of 25 meshes passing the 1 mm 95th-percentile Hausdorff-distance threshold. Task 6 may match either reviewed reference interpretation.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5.5 | 100.0 |
| Gemini 3.8 Flash | 100.0 |
| GPT-5.6 Sol Pro | 100.0 |
| GPT-6 Astra | 100.0 |
| GPT-6 Sol | 100.0 |
| GPT-6 Luna | 96.0 |
| Gemini 3.1 Pro Preview | 92.0 |
| GLM-5.3 Flash | 92.0 |
| GPT-5.6 Luna Pro | 92.0 |
| GPT-5.6 Sol | 92.0 |
| Grok 4.7 | 92.0 |
| Gemini 3.7 Flash | 88.0 |
| GPT-5.6 Luna | 88.0 |
| Kimi K3 | 88.0 |
| Muse Spark 1.3 | 88.0 |
| Claude Fable 5 | 84.0 |
| GLM-5.3 | 84.0 |
| GPT-5.6 Terra Pro | 84.0 |
| Qwen3.7-Max | 84.0 |
| Claude Fable 5.1 | 80.0 |
| Opus 4.8 | 80.0 |
| Claude Opus 5 | 80.0 |
| MiMo-V2.6-Pro | 80.0 |
| Opus 4.6 | 76.0 |
| DeepSeek V4.1 Flash | 76.0 |
| Gemini 3.5 Flash | 76.0 |
| Gemini 3.6 Flash | 76.0 |
| GPT-5.5 | 76.0 |
| Grok 4.5 | 76.0 |
| Grok 4.6 | 76.0 |
| Sonnet 5 | 72.0 |
| Opus 4.7 | 68.0 |
| GPT-5.6 Terra | 68.0 |
| Grok 4.3 | 68.0 |
| MiMo-V2.6-Flash | 68.0 |
| Muse Spark 1.1 | 68.0 |
| Gemini 3 Pro Preview | 64.0 |
| GLM-5.1 | 64.0 |
| Kimi K2.6 | 64.0 |
| Muse Spark 1.2 | 64.0 |