Atlas

Benchmarks

← All benchmarks

CadQueryEval — Component Count

Code · 2026-01-10

Percentage of CadQueryEval tasks whose generated mesh has the expected number of connected components.

Top models (higher is better)

ModelScore
Gemini 3.1 Pro Preview100.0
GPT-5.6 Sol Pro100.0
GPT-5.6 Sol96.0
GPT-5.592.0
GPT-5.6 Luna Pro92.0
GPT-5.6 Terra Pro92.0
Opus 4.888.0
GPT-5.6 Luna88.0
GPT-5.6 Terra88.0
Grok 4.588.0
Kimi K388.0
Qwen3.7-Max88.0
Claude Fable 584.0
Sonnet 584.0
DeepSeek-V4-Pro84.0
Opus 4.580.0
Opus 4.680.0
Gemini 3.5 Flash80.0
Grok 4.380.0
MiniMax M380.0
Opus 4.776.0
Gemini 3 Pro Preview76.0
GLM-5.176.0
GLM-5.276.0
Kimi K2.576.0
Muse Spark 1.176.0
GPT-5.472.0
GPT-5 Mini72.0
Kimi K2.672.0
Qwen3.7-Plus72.0
Gemini 3 Flash Preview68.0
GPT-568.0
GPT-5.168.0
GPT-5.268.0
Opus 464.0
Sonnet 4.664.0
Hy3 preview64.0
O4 Mini64.0
Sonnet 4.560.0
DeepSeek-V4-Flash60.0
Loading Atlas data…