Qwen3.8-27B quantized to run on 12GB RAM; router claims 97.5% of Opus 5.5 at 1/10 cost
Saul_Loveman · x · 2026-11-13
Qwen3.8-27B, quantized by Unsloth under Apache 2.0, now targets setups with just 12GB of RAM. An open-source router also reports achieving 97.5% of Opus 5.5 quality at roughly one-tenth the cost — pushing high-quality inference toward cheap local setups.
More from Infra
- Michael Burry says 'get out' of Nvidia, citing hidden debt and leveraged customer financing — SumitGup · 2026-11-13
- Nvidia's $500B financing backstop with six asset managers has produced zero deals in months — GaryMarcus · 2026-11-13
- After Microsoft outage, open-weight model torrent seeders hit record 30,000 — WolframRvnwlf · 2026-11-13
- AWS disputes outage reports: services 'have been fine', check Health Dashboard — Kyrannio · 2026-11-13
- The math says GPU export controls are working: only a sliver of Nvidia output leaks — bookwormengr · 2026-11-13
- NIDO ships fully offline open-source Android AI assistant with encrypted local RAG and E2EE mesh — NIDO_007 · 2026-11-13