P
roadkill · Production library run 1
29.0Library score
Pass rate75.9%
Simplicity45.6
Problems $$11.56
Implementers use the real production library (uxsim).
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE83.4%51.7$1.78
GLM 5.3 Flash · mini-SWE83.8%38.1$7.69
GPT-5.6 Luna · Codex61.5%46.0$2.08
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean
transit-route-check18.814.21.811.6live-session8.80.02.23.7charging-routes70.10.00.023.4link-time-bins83.460.769.571.2explicit-fleet58.233.360.550.6braess-network31.611.739.427.6signal-plan-sweep10.611.141.721.1merge-spillback45.341.39.532.0incident-guidance10.336.74.717.2corridor-replications60.738.50.033.1speed-limit-control12.712.67.510.9reactive-signals53.723.352.943.3signal-progression24.530.423.826.3gridlock-gating49.144.649.447.7workzone-recovery55.614.334.134.6district-los8.610.88.59.3