heretical · GPT-5.6 Sol run 1

38.2Library score
Pass rate86.6%
Simplicity49.3
Problems $$12.02
Library $$2.26
The library
heretical · designed by GPT-5.6 Sol (Codex)730 source lines
ImplementerPass rateSimplicityCostScore
DeepSeek V4.1 Flash · mini-SWE89.7%50.4$3.14
41.1
GLM 5.3 Flash · mini-SWE79.7%50.9$6.36
34.8
GPT-5.6 Luna · Codex90.5%46.7$2.51
38.9
ProblemDeepSeek V4.1 FlashGLM 5.3 FlashGPT-5.6 LunaMean

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