heretical · DeepSeek V4 Pro run 2

24.9Library score
Pass rate81.3%
Simplicity36.4
Problems $$10.51
Library $$0.21
The library
heretical · designed by DeepSeek V4 Pro (mini-SWE)1,437 source lines
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE86.3%36.0$4.04
27.8
GLM 5.3 Flash · mini-SWE72.6%37.7$3.86
20.0
GPT-5.6 Luna · Codex84.1%35.6$2.61
26.8
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean

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