llamaperf
Submit

Qwen3.8 27B

on 4× NVIDIA RTX 5060 Ti 16GB · llama.cpp

NVIDIA hardware
Sep 7, 2026
Share
Throughput
50-60 t/s gen
VRAM reported
16 GB

Summary

User reports Qwen3.8-27B at 50-60 t/s on 4x RTX 5060 Ti 16GB. Setup is tensor parallelism with MTP. User asks about memory overclocking risks.
View source ↗