Reported in recent model launches
These model reports include this benchmark. Their harnesses, task subsets, scoring policies, and resource budgets can differ; follow each report’s evaluation notes before joining the numbers.
What it measures
Visual-to-code tasks assess whether a model can construct a CAD representation matching a target shape.
How to read the result
Geometric overlap evaluates the resulting object. It should be read alongside validity and the conditions under which code and geometry were generated.
Limits and comparison traps
Recent launches use different subsets and modifications. Tool-enabled and single-shot results, voxel-based overlap, and task counts must stay explicit.
Published evaluation
Exact values from the publisher’s embedded full comparison table. Maximum effort, fallback settings and dated model versions remain in model labels. Blank cells stay missing. Source footnotes specify mixed harnesses and externally quoted comparisons; this is a launch-reported cohort rather than a new independent evaluation. BenchCAD uses the 1,000-file Voxel-IoU subset; tools and no-tools are separate measures.
Where this fits in an evaluation suite
The agent wrapper is part of the system. Pin tools, dependencies, timeouts, tests, and the repository revision. Compare contest coding separately from repository maintenance.
Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.
A reproducible comparison
Exact values from the publisher’s embedded full comparison table. Maximum effort, fallback settings and dated model versions remain in model labels. Blank cells stay missing. Source footnotes specify mixed harnesses and externally quoted comparisons; this is a launch-reported cohort rather than a new independent evaluation. BenchCAD uses the 1,000-file Voxel-IoU subset; tools and no-tools are separate measures.
Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.
PRIMARY EVIDENCE & FURTHER READING
Reviewed 2026-10-27. Explanation and interpretation by RLTasks. Benchmark ownership remains with its original authors.