umami · Kimi K3 run 1
45.8Library score
Pass rate96.2%
Simplicity50.1
Problems $$8.49
Library $$3.44
The library
umami · designed by Kimi K3 (mini-SWE)1,505 source lines
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE96.0%57.0$1.56
GLM 5.3 Flash · mini-SWE96.4%47.2$5.36
GPT-5.6 Luna · Codex96.3%46.0$1.57
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean
Click a problem to see how each solution's score is built.