tin-hat/Qwen-tin

Qwen3.8-Flash-Next (IQ3_S, 84 GB) on 8 GB VRAM and 32 GB RAM: a llama.cpp fork that streams MoE experts from NVMe

1 best points · 0 comments · 1 Hacker News post · 5 GitHub stars · View on GitHub

Hacker News discussions

• Run Qwen3.8-Flash-Next on 8GB VRAM and 32 GB RAM at 20T/S (1 points, 0 comments, by chess39, 2026-10-19)