P

arazu · Production library run 1

40.1Library score
Pass rate76.0%
Simplicity66.9
Problems $$26.22

Implementers use the real production library (rapier).

ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE82.6%63.4$3.62
42.8
GLM 5.3 Flash · mini-SWE73.3%66.6$16.93
39.0
GPT-5.6 Luna · Codex72.1%70.7$5.66
38.6
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean
lander-lidar43.143.943.043.3scene-contact50.149.229.643.0world-contact79.912.332.041.4aircraft-raycast42.262.237.447.3capsule-training26.70.228.918.6physics-tick10.23.96.56.9offset-strike53.570.356.960.2layered-laser81.267.985.878.3swept-cargo38.046.835.840.2compound-scanner25.174.359.252.9sensor-transit-ledger41.832.767.747.4articulated-creature36.40.630.522.5cable-replay33.743.10.025.6damping45.846.750.247.6one-way34.928.739.534.4material-braking34.342.940.139.1platformer-ccd36.742.421.133.4grab-replay73.349.054.959.1keyed-platform-drop58.331.645.945.3ccd-wall3.322.919.115.1motor-target30.57.30.012.6rope-cut62.978.664.768.7