Hacker News new | past | comments | ask | show | jobs | submit
'Good' news, there seems to be nothing new architecture-wise. Same as Qwen 3.5 and 3.6, so llama.cpp doesn't know the difference.