LINKED LIST [txt mode] ▸ Qwen 3.8 27B is excellent, but it def...

home explore | log in

Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things

simonwillison.net · first added by @hnl · 2026-09-25 · 2 upvotes

log in to save, upvote or flag this.

─── In 1 list ──────────────────────────────────────────

* Issue #806 by @hnl [favorites]

* Best value LLM (from the discussion)

─── Discussions ────────────────────────────────────────

* Qwen 3.8 27B is excellent, but it defaults to overthinking things

802 pts · 385 comments · node

─── From the discussion ────────────────────────────────

* Qwen/Qwen3.8-27B-FP8 · Hugging Face

huggingface.co · node

* pelican-box.md

gist.github.com · node

* froggeric/Qwen-Fixed-Chat-Templates · Hugging Face

huggingface.co · node

* GitHub - laurencehardman/llama-mindcontrol: LLM inference in C/C++

github.com · node

gist.github.com · node

─── Also saved alongside this ──────────────────────────

* License Plate Reader Searches Should Require a Warrant

3 upvotes · 2026-08-13

* Guillaume Técher

2 upvotes · 2026-08-13

* Why Go is an Ideal Language for AI-Assisted Software Engineering- Google Developers Blog

2 upvotes · 2026-08-13

* The August 17 outage, and the work ahead

4 upvotes · 2026-09-13