DeepSeek V4 Flash 284B (13B active)
on NVIDIA RTX 5090 · llama.cpp · 1,048,576 ctx
Sep 7, 2026
Summary
User reports prefill of ~650-700 t/s and decode of ~17 t/s.
Setup uses CPU offloading for MoE experts, with a loading time of 32 sec.
User notes the speed is not as impressive as Qwen models but sees room for optimization.