–

weft · No library run 1

34.3Library score
Pass rate82.2%
Simplicity49.0
Problems $$4.05

Implementers get no library and write everything from scratch.

ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE91.9%45.7$1.07
38.9
GLM 5.3 Flash · mini-SWE68.7%54.3$1.40
28.2
GPT-5.6 Luna · Codex86.2%47.1$1.58
35.8
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean
bag-rules41.032.317.430.2ninja-build26.638.739.534.9mail-addresses54.40.058.037.5property-paths51.153.450.251.6formula-ledger31.731.225.429.4record-spans33.639.745.939.7block-events47.823.945.038.9command-errors23.331.119.424.6route-tags54.30.146.733.7template-audit35.635.630.633.9grammar-inventory34.818.422.025.1grammar-trace48.146.844.546.5nesting-guard26.220.522.323.0unicode-spans36.723.034.831.5