IT
Prenoti una chiamata

Local AI · Recipe · 4× DGX Spark

Qwen3.8-Flash-Next 125B NVFP4 with vLLM on 4× DGX Spark

Qwen3.8-Flash-Next 125B NVFP4 with vLLM on 4× DGX Spark: 40.5 tok/s according to github.com (dataset as of Sep 29, 2026).

by tonyd2wild

40.5tok/sMixed

1 request, mixed prompt set, with MTP (per recipe)

Source: github.com
Intelligence (original model) · Artificial Analysis40with thinking
Instructions in the README Repo

Engine

Engine
vLLM nightly 8a728663 + PLE-mmap-Patch
Quantization
NVFP4 (nvidia), FP8 KV
Model family
Qwen3.8-Flash-Next
Context
262,144
Parameters
125B-A6B
Creator
tonyd2wild
GitHub stars
125
Repo updated
Sep 7, 2026

Measurements

Every sourced value of this recipe, each with its condition and source. Bars relative to the largest value in the group.

Decode · 1 request

  • Mixed (decode)40.5 tok/sSingle stream, median of 40 prompts; TTFT 220 ms Source 
  • Peak (decode)40.5 tok/sSingle stream, median of 40 prompts; TTFT 220 ms Source 

Prefill by context

What you need

Hardware
4 × NVIDIA DGX Spark (GB10)
Weights
Engine
vLLM nightly 8a728663 + PLE-mmap-Patch
Context
262,144 tokens

Notes

What matters before you rebuild it.

  • Custom kernel required
  • License unclear

Sources

Related recipes

← Back to overview

Local AI in your company?

In a workshop we work out which models and which hardware fit your tasks, and build the first agent on your infrastructure.

See the AI workshopBack to overview