BenchCAD 1.0 - Code QA: leaderboard

From a part's CadQuery source code, the model answers numeric geometry questions (ratios within 5%, integers exactly) across 2,400 questions; the score is the share answered correctly. It asks the same questions as BenchCAD 1.0 Vision QA, which shows rendered images instead of code.

Metric: Total accuracy (0 to 1) on 2,400 numeric CAD questions from CadQuery source. Source: benchcad.com. Saturation forecast: Around February 2027. 8 models tracked. Updated 16 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3.1 Pro (Preview)Thinking0.838#93
2Claude Opus 4.7Thinking0.829#78
3GPT-5.30.823#197
4GPT-4o0.726#744
5O30.708#228
6GPT-OSS-120B0.689#703
7Nemotron 3 Super0.671#775
8Gemma 4 31B (IT)0.664#350

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=benchcad-1-0-code-qa · How It Works · Data refreshed daily, snapshot 2026-11-16.