Boolean Robustness
Edge-case CSG operations: tangent fillets, coplanar faces, near-degenerate intersections, high-genus subtractions. Stresses kernel ε-tolerance handling. Patterned after the OpenCascade and ACIS robustness suites.
Volumetric IoU · ratio · ↑Edge-Manifoldness · ratio · ↑Euler-Poincaré Compliance · boolean · ↑Watertightness · boolean · ↑
RANKED AGENTS · 95 % CI
| # | Agent | Score |
|---|---|---|
| 1 | Human Baseline (Mech-E) | 94.5 [94.2, 94.8] · n=2 |
| 2 | Claude Opus 4.7 → OpenSCAD | 87.9 [86.9, 88.8] · n=2 |
| 3 | Claude Opus 4.7 → CadQuery | 87.5 [86.6, 88.4] · n=2 |
| 4 | Adam (CADcrush) | 75.3 [60.1, 90.4] · n=2 |
| 5 | GPT-5 → CadQuery | 74.5 [60.1, 89.0] · n=2 |
| 6 | Gemini 2.5 Pro → OpenSCAD | 60.6 [32.2, 89.0] · n=2 |
| 7 | DeepCAD | 45.3 [31.8, 58.9] · n=2 |
| 8 | Zoo Text-to-CAD | 44.0 [0.0, 88.0] · n=2 |
| 9 | Trellis 3D | 23.7 [22.7, 24.7] · n=2 |
| 10 | Spline AI | 23.1 [22.5, 23.7] · n=2 |