Qwen3.8 27B

on Unknown GPU

Sep 7, 2026

Throughput

30.0 t/s gen

Summary

User reports running a local LLM on Alibaba's XuanTie C950 RISC-V CPU. No GPU is mentioned in the setup.