Hands-on: Qwen3.8-27B Q6 Delivers Strong Agentic Coding Performance
Ok_Ninja7526 · reddit · 2026-08-22
The author shares experience using Qwen3.8-27B Q6 for nearly 20 hours of continuous, goal-oriented coding tasks. Running on RTX 3090 and 3060 GPUs, the model maintained a consistent generation speed of 60–63 tokens/s throughout the session, demonstrating robust performance.
More from Models
- Claude Opus 5.5 effort levels tested: low reasoning iterates 9x faster than max — tom_doerr · 2026-10-16
- Matt Turck podcast unpacks Jev: TypeSafe's machine-native AI and its RLHF critique — mattturck · 2026-10-16
- Muse Spark Open runs locally on a Mac M3 Ultra via llama.cpp: frontier model on desktop — pcuenq · 2026-10-16
- OpenAI ships GPT-Live-1.1: new voices in six languages, fewer confirmations — OpenAI · 2026-10-16
- Muse Spark model appears on LMArena, testable in Battle and Agent modes — arena · 2026-10-16
- Running a multi-model UI design arena: Paseo orchestrates four design drafts per model — vista8 · 2026-10-16