plaid · Kimi K3 run 1

38.6Library score
Pass rate95.0%
Simplicity49.3
Problems $$12.94
Library $$4.62
The library
plaid · designed by Kimi K3 (mini-SWE)1,699 source lines
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE98.6%47.8$2.43
46.5
GLM 5.3 Flash · mini-SWE95.9%51.5$8.25
28.7
GPT-5.6 Luna · Codex90.9%49.5$2.26
40.7
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean
ledger-balance39.30.055.031.4typed-settings50.256.246.250.8keymap-resolver44.442.943.743.6template-trace44.20.041.728.6recipe-graph47.559.046.250.9scoped-eval53.70.00.117.9temporal-normalizer45.547.444.345.7script-recovery49.00.031.726.9stack-audit41.341.538.240.3query-resolution60.60.058.439.7selection-flattener46.348.846.347.1schema-resolver49.154.748.550.8grammar-crossref49.928.747.742.1config-digest38.60.028.322.3shape-atlas38.251.133.941.0