canon · Kimi K3 run 3
56.3Library score
Pass rate94.0%
Simplicity63.6
Problems $$6.15
Library $$5.61
The library
canon · designed by Kimi K3 (mini-SWE)4,032 source lines
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE97.5%63.7$1.96
GLM 5.3 Flash · mini-SWE89.9%70.0$2.68
GPT-5.6 Luna · Codex94.5%57.1$1.51
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean
profile-patch72.265.356.964.8thermostat-codec52.314.340.635.7tagged-config62.869.757.763.4rtl-pin-lock91.194.583.789.8pipeline-provenance47.844.130.940.9boundary-views56.847.551.351.9ml-run-config89.198.175.387.5ha-triggers72.956.746.658.7env-config52.872.061.062.0paper-poster49.057.041.149.0typed-client50.632.244.042.3cli-dispatch34.439.829.434.5trading-events73.873.968.071.9llm-contracts38.939.927.235.3