Qwen3.8 27B-heretic text-only MLX 4-bit, single 68-prompt/120-output run

User reports Qwen3.8 at 37.9 t/s on an M3 Ultra 512GB, from a single run of one 68-token prompt and 120 generated tokens. Setup is MLX with affine 4-bit quantization, group size 64 and 4.501 effective bpw, peaking at 15.5 GB of memory. The author describes the figure as a rough guide rather than a repeated controlled benchmark; no prompt-processing throughput or TTFT is given.