CadEval
CadEval · evaluates the ability to generate and reason about Computer-Aided Design code, testing spatial reasoning and engineering knowledge.
The Frontier
Best score over time · one chart, every benchmark
Full rankings
17 models tested · sorted by score
| # | Model | Score |
|---|---|---|
| 1 | 74.0 | |
| 2 | 64.0 | |
| 3 | 62.0 | |
| 4 | 56.0 | |
| 5 | 54.0 | |
| 6 | 54.0 | |
| 7 | 48.0 | |
| 8 | 42.0 | |
| 9 | 34.0 | |
| 10 | 34.0 | |
| 11 | 32.0 | |
| 12 | 30.0 | |
| 13 | 30.0 | |
| 14 | 26.0 | |
| 15 | 26.0 | |
| 16 | 16.0 | |
| 17 | 12.0 |
Score distribution
Where models cluster
Correlated benchmarks
Pearson r · original research
Benchmarks that track with CadEval
Pearson correlation across models scored on both benchmarks. Closer to 1 = strongly predictive.
Frequently asked
About CadEval
What does CadEval measure?
CadEval · evaluates the ability to generate and reason about Computer-Aided Design code, testing spatial reasoning and engineering knowledge. 17 AI models have been tested on it. Scores range from 12.0 to 74.0 out of 100.
Which model leads on CadEval?
o3 from OpenAI leads CadEval with a score of 74.0. The median score across 17 tested models is 34.0.
Is CadEval saturated?
No · the top score is 74.0 out of 100 (74%). There is still meaningful room for improvement on CadEval.
Does CadEval predict performance on other benchmarks?
Yes · CadEval scores correlate 0.95 with Aider polyglot across 14 shared models. Models that do well on CadEval tend to do well on Aider polyglot.
How often is CadEval data refreshed?
BenchGecko pulls updates daily. New model scores on CadEval appear as soon as they are published by Epoch AI or the model provider.
- Category
- Code
- Max score
- 100
- Models
- 17
- Updated
- 2025-06-17
More code benchmarks
Same category · related evaluations