P
umami · Production library run 3
44.2Library score
Pass rate93.3%
Simplicity51.8
Problems $$10.13
Implementers use the real production library (unagi-chan).
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE93.9%55.7$1.25
GLM 5.3 Flash · mini-SWE93.7%50.6$7.10
GPT-5.6 Luna · Codex92.3%49.1$1.78
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean
Click a problem to see how each solution's score is built.