Kenny.eth
Kenny.eth|Aug 15, 2026 02:44
QWEN3.8-27B brings the GPT-3 moment to local large models. Time to build your own local large model setup! GPU: RTX 3090 24GB CPU: AMD Ryzen 7 7700 / 9700X Board: Mini-ITX B650 RAM: 64GB DDR5 6000 (starting at 64GB) SSD: 2TB PCIe 4.0 NVMe Based on Qwen3.8-27B Dense Q4_K_M / UD-Q4_K_XL, you can achieve speeds of around 40~50 tokens/s. Pretty satisfied with that!
+6
Mentioned
Share To

Timeline

HotFlash

APP

X

Telegram

Facebook

Reddit

CopyLink

Hot Reads