llamaperf

Qwen3.8 125B (6B active) Flash-Next

on M2 Pro 16GB · Strata · 128,000 ctx

Tone: positiveApple hardware
Nov 6, 2026
Throughput
5.5 t/s gen · 77.0 tokens/s prompt processing (prefill)Prompt processing is how fast the input is read, not how long you wait for the first word. How to measure time to first token.
Quant
iQ3_xxs (GGUF)
System RAM
16 GB

Use cases

codingsummarization

Summary

User reports Qwen3.8 Flash-Next at a median 77 tok/s prefill and 5.5 tok/s decode on a 16GB M2 Pro Mac Mini. Setup is Strata with the iQ3_xxs quant at up to 128k context; experts are swapped across prompts, so the user assumes speeds reflect worst-case full-context behavior. The medians span tests from 16k to 120k context. The user calls it a barely passable chat generator and a last-resort code helper, and notes heavy SSD write load.

Compare this run

Medians of plain runs: one device, one request, no speculative decoding, the whole model in the device's memory. The report above is one setup, so read it beside them and not as one of them.

Qwen3.8 125B · 6B active at 4-bit by hardware

Bars are medians on one scale, 0 to 232.0 tokens/s. Each dot is one run and opens its report. Press ▶ to watch text arrive at a row's speed.

Everything reported on the M2 Pro 16GB →Every Qwen3.8 report →Estimate Qwen3.8 on this card →