Tag archive
Qwen 3.8 27B is a free open-weight multimodal model that runs locally on a single GPU. Here is exactly how to set it up with Ollama, llama.cpp, or vLL
Qwen3.8-Flash-Next needs only 6B active parameters per token, so the internet decided it runs on 12GB of VRAM. It does — but only if you also have 75GB of total memory and an SSD willing to stream a 5
Qwen3.8-27B is the first Apache-2.0 model that scores 61.7 on SWE-bench Pro and still fits on one 24GB GPU. Here is the working-developer build: which of the 790 GGUF quants to actually download (with
Originally published at norvik.tech Introduction Deep dive into Qwen 3.8 27B release,...
Agentic coding on a Strix Halo laptop with Qwen3.8-27B and Flash-Next, the LlamaStash and Pi setup behind it, and the tuning
Day-one measurements of qwen3.8-max at $2/$6: an effort dial that is really a budget cap, a 1/6 accuracy collapse with thinking off, and a 16-token fix.
Originally published at norvik.tech Introduction Explore the technical aspects of...
AI Dev Weekly is a Thursday series where I cover the week's most important AI developer news, with my...