Qwen3.8 125B (6B active) Flash-Next
on M3 Ultra 96GB · ds4 · 262,144 ctx
Use cases
coding
Summary
User reports Qwen3.8 Flash Next Q4 running through ds4 on an M3 Ultra 96GB with the full 262K context configured, delivering 55–60 tok/s decode and 667 tok/s prefill.
About 80GB is used for model, KV cache and buffers, while the model's 95GB n-gram table streams from SSD.
The user runs it daily through a codex harness as a headless box.