Atlas

Benchmarks

← All benchmarks

CadQueryEval Corrected Geometry Volume

Code · 2026-09-30

CadQueryEval final corrected-geometry release: percentage of 25 meshes passing the 2 percent volume tolerance. Task 6 may match either reviewed reference interpretation.

Top models (higher is better)

ModelScore
Claude Opus 5.5100.0
GPT-5.6 Sol Pro100.0
GPT-6 Astra100.0
GPT-6 Luna100.0
GPT-6 Sol100.0
Gemini 3.8 Flash96.0
GLM-5.3 Flash92.0
Gemini 3.1 Pro Preview88.0
GPT-5.6 Luna88.0
GPT-5.6 Luna Pro88.0
GPT-5.6 Sol88.0
GPT-5.6 Terra Pro88.0
Grok 4.788.0
Gemini 3.7 Flash84.0
Kimi K384.0
Muse Spark 1.384.0
Claude Fable 580.0
Claude Fable 5.180.0
Claude Opus 580.0
DeepSeek V4.1 Flash80.0
Gemini 3.5 Flash80.0
GLM-5.380.0
Qwen3.7-Max80.0
GPT-5.576.0
Grok 4.676.0
MiMo-V2.6-Pro76.0
Opus 4.872.0
Sonnet 572.0
Gemini 3.6 Flash72.0
Grok 4.372.0
Grok 4.572.0
MiMo-V2.6-Flash72.0
Opus 4.668.0
GPT-5.6 Terra68.0
GLM-5.164.0
Muse Spark 1.164.0
Opus 4.760.0
Gemini 3 Flash Preview60.0
Gemini 3 Pro Preview60.0
Kimi K2.660.0
Loading Atlas data…