Gufo-Qwen3.6-35B-A3B-Q6dense – 3095tok/s prefill; 190 tok/s decode on Strix Halo | Hacker News - But Better

Gufo-Qwen3.6-35B-A3B-Q6dense – 3095tok/s prefill; 190 tok/s decode on Strix Halo github.com

3000 tok/s prefill; 120 tok/s decode on Strix Halo. This is a fork of gufo to support Qwen 3.6 35B A3B for workloads that prioritizes speed over intelligence. - NinjaPear/gufo-Qwen3.6-35B-A3B-Q6dense

1 points•nubela•about 3 hours ago•0 comments•

#gufo #tok #35b #a3b #35b a3b #q6dense #prefill #qwen3

0 comments

No comments yet.

Related stories

MoE Analysis Qwen[3.5|3.6]35B-A3Bgslaller.github.io
Hacker News · 1 points · 4 days ago
Launch HN: Magnitude (YC S25) – Self-optimizing inference engine for agentsgithub.com
Hacker News · 186 points · 3 days ago
Prefill Concurrency in SGLang: Consistent TTFT Under Multi-Tenant Loadsference.com
Hacker News · 1 points · 7 days ago
Halogen – Fastest Way to Run Qwen3.8-Flash-Next on AMD Strix Halogithub.com
Hacker News · 1 points · 3 days ago
ThinkingCap-Qwen3.8-27B: the same answers, 37% less thinkingbottlecapai.com
Hacker News · 2 points · 9 days ago
Kev: Tiny Jev-like family of decision models built on top of Qwen3.5github.com
Hacker News · 455 points · 12 days ago