–
plaid · No library run 2
35.6Library score
Pass rate94.5%
Simplicity41.5
Problems $$9.01
Implementers get no library and write everything from scratch.
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE98.4%40.9$1.41
GLM 5.3 Flash · mini-SWE93.3%43.6$5.47
GPT-5.6 Luna · Codex91.5%40.3$2.13
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean
ledger-balance39.238.136.537.9typed-settings44.10.029.324.5keymap-resolver41.442.932.739.0template-trace50.548.540.746.6recipe-graph48.256.352.052.1scoped-eval40.80.037.025.9temporal-normalizer29.119.718.422.4script-recovery37.333.731.534.2stack-audit37.536.734.636.3query-resolution45.442.817.335.1selection-flattener39.227.142.136.1schema-resolver40.142.237.239.8grammar-crossref37.044.634.738.8config-digest27.624.526.826.3shape-atlas37.140.838.138.7