heretical · DeepSeek V4 Pro run 1

33.7Library score
Pass rate89.0%
Simplicity42.3
Problems $$10.66
Library $$0.26
The library
heretical · designed by DeepSeek V4 Pro (mini-SWE)1,977 source lines
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE90.5%39.9$4.39
33.0
GLM 5.3 Flash · mini-SWE88.4%48.5$3.70
37.5
GPT-5.6 Luna · Codex88.1%38.6$2.58
30.6
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean

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