canon · Kimi K3 run 1

50.7Library score
Pass rate92.0%
Simplicity59.0
Problems $$6.43
Library $$7.07
The library
canon · designed by Kimi K3 (mini-SWE)3,922 source lines
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE95.1%58.5$2.30
52.6
GLM 5.3 Flash · mini-SWE93.1%64.3$2.61
55.3
GPT-5.6 Luna · Codex87.8%54.1$1.52
44.3
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean
profile-patch46.747.30.731.6thermostat-codec46.655.545.549.2tagged-config54.450.849.951.7rtl-pin-lock85.796.381.187.7pipeline-provenance58.769.442.056.7boundary-views41.845.130.839.2ml-run-config84.182.883.983.6ha-triggers54.856.644.051.8env-config51.350.855.352.5paper-poster39.950.138.542.8typed-client44.550.820.638.6cli-dispatch35.834.430.333.5trading-events55.537.257.049.9llm-contracts36.546.940.541.3