Four Open-Weight Models Worth Running Locally, from 27B to 2.8T

drdanielbender's thread recommends four open-weight models (27B to 2.8T parameters), arguing the best model is one you can legally run locally: Kimi K3.1 Pro tops the intelligence index at 46, while Qwen3.8-27B is the strongest small model, fitting in 24GB VRAM with Q4 quantization.

2026-11-13 ~ 2026-11-13 · 4 related posts

• Open-weight shortlist from Kimi 2.8T down to Qwen 27B: practical beats smartest — drdanielbender · 2026-11-13
• Kimi tops open-weight intelligence index at 46, but 2.8T-param model is the priciest to run — drdanielbender · 2026-11-13
• Qwen3.8-27B hits 20 points: strongest small model, fits a 24GB GPU at Q4 — drdanielbender · 2026-11-13
• Qwen3.8-27B called strongest small model: fits a 24GB GPU at Q4 quantization — drdanielbender · 2026-11-13