Qwen 3.8 27B Feels Just 'OK' in Local Testing Despite Hype
YeetHub · reddit · 2026-10-06
A professional developer ran Unsloth's Q6 quant of Qwen 3.8 27B on 32GB VRAM with OpenCode: 30-40 t/s decode, fine on small tasks, but anything with nuance sends it into 'but wait/actually' thinking loops that can eat the entire 95k context with nothing done. He sees a gap between online sentiment ('as good as Opus 4.5') and local reality, and questions industry decisions if this is what vibe coding runs on.
More from Models
- Dev bets Thinking Machines' "fledge alpha" will ship as "fledgling" — willcb · 2026-10-06
- Insider speculation: GPT-6.5 expected within 4 weeks, above Fable 5.5 level — VraserX · 2026-10-06
- Gemini users hit image generation limits as Google tightens daily quotas — BrattyMiku · 2026-10-06
- Indie dev releases Ekbasis, a 27B world model, weights live on Hugging Face — Over_Monitor_8770 · 2026-10-06
- OpenRouter's top 10 providers by monthly token consumption reveal real usage, not benchmarks — FinanceYF5 · 2026-10-06
- "There's no reason for models to struggle like this": debate over AI naming chaos — a1zhang · 2026-10-06