Qwen3.5 27B
on NVIDIA RTX 5090 · llama.cpp · 1,024 ctx
Summary
User benchmarks Qwen3.5-27B Q4_K_M on an RTX 5090 at 30.09 t/s generation and 1515.83 t/s prompt processing with llama-bench (pp1024/tg400, 5 reps).
Setup is llama.cpp b10741 with CUDA 13.3 on Windows 11, -ngl 99 and flash attention on, 15.65 GiB GGUF.
A Qwen3 MoE 30B-A3B control on the same rig reaches 134.18 t/s tg, ruling out a global Windows/driver overhead. The user notes native Linux with MTP reaches about 76% of roofline and ~1.7x speedup, while the Windows/Ollama path runs 1.5-1.6x slower at every draft depth. PCIe Gen5 x1 vs x4 gave 27.95 vs 30.09 t/s, so decode is not host-link-bound.