Arao AI
Back to search results

Qwen3.5-9B · Q8_0

Start a discussion
NVIDIA DGX Spark128GBvLLM v0.52L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-9B (Q8_0) on NVIDIA DGX Spark with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

101.73 tok/s

Decode

Decode speed

523.44 tok/s

Prefill

Prefill speed

2.3 s

TTFT

Time to first token

23.84 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

  1. llama.cpp · Q6_K Decode 110.08 · Prefill 496.59 ·
  2. Strata · Q3_K_M Decode 209.28 · Prefill 1036.76 ·
  3. llama.cpp · IQ2_M Decode 181.94 · Prefill 684.54 ·
  4. vLLM · IQ2_M Decode 210.69 · Prefill 1114.84 ·
  5. vLLM · Q8_0 (current)Decode 101.73 · Prefill 523.44 ·

Core figures

Decode (avg)
101.73 tok/s
Prefill (avg)
523.44 tok/s
TTFT (avg)
2.3 s
VRAM (avg)
23.84 GB
MTP acceptance rate
—
TTFB
0.81 GB
Power draw
239.7 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-9B
Quantization
Q8_0
Framework
vLLM
Version
v0.52
Context length
16384 tokens
Batch size
16
GPU layers
999
Flash Attention
On

Hardware

Nominal and measured figures are shown side by side; whether it runs is the reader's call

GPU
NVIDIA DGX Spark
Nominal VRAM
128 GB
Measured VRAM (avg)
23.84 GB
OS
Ubuntu 24.04
Driver
565.57
CUDA
12.8
Power draw
239.7 W

Sources & evidence

1 measured records in total, each traceable to its original source

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-13

    v0.52 · Ubuntu 24.04 · CUDA 12.8 · 16384 ctx

    101.73 tok/s

    Decode

    523.44 tok/s

    Prefill

    2.3 s

    TTFT

    23.84 GB

    VRAM

    —

    MTP

    239.7 W

    Power draw

    L2 batch seed 2026-10-13: vLLM on hw_id=54 running model_id=49 at Q8_0