Qwen 3.5 is coming
Looks to be 35b MoE and 9b dense
https://github.com/ggml-org/llama.cpp/pull/19468Open linkView original on lemdro.idSyndicated from the fediverse. Read and engage on the original instance.
View original on lemdro.idhttps://github.com/ggml-org/llama.cpp/pull/19468Open linkView original on lemdro.id
2 replies
Holy crap they have momentum !
Qwen has been the only models at such sizes with competence in programming, and so big deal.