DeepSeek V4 Flash 284B (13B active)
on M3 Ultra 512GB · llama.cpp · 10,000 ctx
Sep 7, 2026
Summary
User reports 8 t/s on a Mac Studio M3 Ultra 512GB with DeepSeek V4 Flash GGUF Q4_K_XL.
The user also tried Q8.
The user expected better performance.
on M3 Ultra 512GB · llama.cpp · 10,000 ctx