weft · GPT-5.6 Sol run 2

31.4Library score
Pass rate80.3%
Simplicity46.5
Problems $$12.74
Library $$1.92
The library
weft · designed by GPT-5.6 Sol (Codex)2,241 source lines
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE86.6%46.6$2.58
36.5
GLM 5.3 Flash · mini-SWE71.5%48.1$7.85
27.0
GPT-5.6 Luna · Codex82.2%44.9$2.31
30.8
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean
bag-rules32.935.026.731.5ninja-build5.523.522.017.0mail-addresses40.136.747.641.5property-paths51.832.753.546.0formula-ledger31.60.029.720.4record-spans61.848.924.044.9block-events44.852.344.147.0command-errors24.333.825.727.9route-tags45.828.732.635.7template-audit42.228.534.535.1grammar-inventory33.333.322.329.6grammar-trace44.50.418.921.3nesting-guard23.30.015.312.9unicode-spans29.324.834.429.5