simu · Kimi K3 run 3
35.9Library score
Pass rate87.1%
Simplicity55.8
Problems $$4.03
Library $$2.57
The library
simu · designed by Kimi K3 (mini-SWE)1,134 source lines
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE91.0%59.1$1.57
GLM 5.3 Flash · mini-SWE79.2%58.7$1.20
GPT-5.6 Luna · Codex86.1%51.5$1.25
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean
intermodal-terminal-flow29.20.047.525.6diagnostic-specimen-release54.40.023.425.9rail-corridor-recovery96.50.069.355.2binary-offload-episode57.50.064.840.8packet-schedulers45.30.026.824.0wildfire-relief-depot54.20.039.431.2preempted-fault-race71.675.559.969.0rendezvous-rollback51.00.042.131.0filter-reservation31.949.424.635.3one-shot-bell55.958.448.754.3quorum-poisoned-losers44.754.037.445.4harbor-surge26.80.010.112.3harbor-vessels22.10.029.217.1