Antigravity 2.0 Tops the OpenSCAD Architectural 3D LLM Benchmark
The OpenSCAD Architectural 3D LLM Benchmark, hosted on ModelRift, evaluates large language models on their ability to generate valid OpenSCAD code for architectural 3D models. Antigravity 2.0 achieved a 72.4% success rate, surpassing competitors like GPT-4o (68.1%) and Claude 3.5 Sonnet (65.3%). The benchmark includes 200 tasks ranging from simple shapes to complex structures like staircases and roofs. Antigravity 2.0's performance suggests improved spatial reasoning and code generation capabilities, which could streamline 3D modeling workflows for architects and developers. The model is available via API with pricing at $0.15 per 1K input tokens and $0.60 per 1K output tokens.
Developers can now generate accurate 3D architectural models from text, reducing manual coding effort.