Skip to content

Benchmark

BenchCAD · Vision2Code voxel IoU

Higher scores are better.

Models measured
8
Scores recorded
10
Sources
1
IQ weight
No weight published

Score distribution

0.464
Mean
0.497
Median
0.204
Std dev
0.266
10th percentile
0.73
90th percentile
0.155
Min
0.1550.747
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Claude Sonnet 5.5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted · 1,000-file subset · no tools · max · 5 runs· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.747
2Claude Opus 5.5Opus 5.5 System Card capability section · 1,000-file subset · 256×256 renders · no tools · max · 5 runs· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.73
3Claude Haiku 5.5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted · 1,000-file subset · no tools · max · 5 runs· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.67
4Claude Fable 5.1Opus 5.5 System Card capability section · 1,000-file subset · 256×256 renders · no tools· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.606
5Claude Opus 5Opus 5.5 System Card capability section · 1,000-file subset · 256×256 renders · no tools· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.497
6Claude Sonnet 51,000-file subset · Python + image crop · 5 runs· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.373
7Claude Sonnet 4.6Python + image crop · Sonnet 5 card rerun· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.327
8Claude Sonnet 4.61,000-file subset · no tools · Sonnet 5 card rerun· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.267
9Claude Sonnet 51,000-file subset · no tools · 5 runs· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.266
10Claude Haiku 4.5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted · 1,000-file subset · no tools · 5 runs· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.0.155

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.