Benchmark Radar™
Star —

BenchCAD

Explore score history ↗

BenchCAD is a benchmark for programmatic CAD reasoning built from 17,900 execution-verified CadQuery programs spanning 106 industrial part families, roughly half anchored to real ISO, DIN, EN, ASME, and IEC specification tables. It decomposes CAD capability into matched tasks; the Vision2Code task requires models to generate CadQuery code from multi-view renders, scored by voxel IoU against the reference geometry.

Modality
multimodal
Evaluates
multimodal, reasoning, code, vision
Openness
unknown
Importer
LLM Stats

Source provenance

BenchCAD results and reported scores

Reported benchmark scores · LLM Stats

Reported scores: 4

ModelOrganizationReported valueReportedEvidence
Claude Sonnet 5Anthropic0.3732026-06-30evidence
GPT-5.6 LunaOpenAI0.6312026-07-09evidence
GPT-5.6 SolOpenAI0.7062026-07-09evidence
GPT-5.6 TerraOpenAI0.6232026-07-09evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.

Which sources cite BenchCAD?

Related multimodal, reasoning, code, vision benchmarks