---
type: "Article"
title: "Qwen3.8-27B NVFP4 with vLLM on 1× DGX Spark (Weschera)"
description: "Qwen3.8-27B NVFP4 with vLLM on 1× DGX Spark (Weschera): recipe by Weschera with speed, condition and source for every number, weights and license."
resource: "https://www.contextstudios.ai/local-ai/weschera--qwen38-27b-nvfp4-unsloth-mtp3-1x-dgx-spark"
language: "en"
generated:
  by: "process:contextstudios-md/1"
  at: "2026-10-02T22:00:09.038Z"
status: "stable"
---

# Qwen3.8-27B NVFP4 with vLLM on 1× DGX Spark (Weschera)

Qwen3.8-27B NVFP4 with vLLM on 1× DGX Spark (Weschera): 45 tok/s according to github.com (dataset as of Sep 29, 2026).

- Creator: [Weschera](https://www.contextstudios.ai/local-ai/creators/weschera.md)
- Engine: vLLM
- Quantization: NVFP4 (unsloth; compressed-tensors) + fp8 KV + triton_attn (fp8-KV auf GB10 nur mit triton, FA kann kein fp8-KV auf sm_121)
- Model family: Qwen3.8-27B
- Hardware: dgx-spark x1
- Context: 262144

## All numbers

- 45 tok/s (Everyday; Everyday, 1 request, prose prompt, with MTP) — [source](https://github.com/Weschera/Qwen3.8-27B-DGX-Spark-Quant-Ladder/blob/main/README.md)
- decode_tps: 45 (1 request) — [source](https://github.com/Weschera/Qwen3.8-27B-DGX-Spark-Quant-Ladder/blob/main/README.md)
- decode_tps: 45 (1 request) — [source](https://github.com/Weschera/Qwen3.8-27B-DGX-Spark-Quant-Ladder/blob/main/README.md)

## Weights

- [unsloth/Qwen3.8-27B-NVFP4](https://huggingface.co/unsloth/Qwen3.8-27B-NVFP4) (main, apache-2.0)
- [Qwen/Qwen3.8-27B-FP8](https://huggingface.co/Qwen/Qwen3.8-27B-FP8) (base, apache-2.0)

[https://github.com/Weschera/Qwen3.8-27B-DGX-Spark-Quant-Ladder](https://github.com/Weschera/Qwen3.8-27B-DGX-Spark-Quant-Ladder)

## Related

- [Local AI on your hardware.](https://www.contextstudios.ai/local-ai.md)