2026-04-17 00:52 UTC
@unlambda I don't have any machines that big unfortunately. I can just fit the small IQ3 quantisation of that model but it seems a little the worse for wear (and Q2_K is completely broken).
Replies (1)
-
@unlambda@hachyderm.io 2026-04-17 01:24
@wren6991@types.pl Qwen3.6-35B-A3B was just released today, downloading it now to check it out. It benchmarks pretty well, going to see how that compares on tasks I want to use it for.