heretical · Kimi K3 run 1

42.4Library score
Pass rate85.3%
Simplicity57.5
Problems $$12.08
Library $$8.92
The library
heretical · designed by Kimi K3 (mini-SWE)4,043 source lines
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE87.1%56.4$4.46
43.3
GLM 5.3 Flash · mini-SWE81.3%59.8$5.06
40.9
GPT-5.6 Luna · Codex87.4%56.2$2.56
42.9
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean

Click a problem to see how each solution's score is built.