Arao AI
Back to search results

Qwen3.5-27B · Q8_0

Start a discussion
NVIDIA H2097GBvLLM v0.41L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-27B (Q8_0) on NVIDIA H20 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

27.55 tok/s

Decode

Decode speed

97.34 tok/s

Prefill

Prefill speed

0.47 s

TTFT

Time to first token

67.3 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

16 days ago

Last verified

Performance

  1. vLLM · IQ2_XS Decode 63.17 · Prefill 260.61 ·
  2. vLLM · Q8_0 (current)Decode 27.55 · Prefill 97.34 ·
  3. Strata · IQ3_S Decode 67.46 · Prefill 225.71 ·

Core figures

Decode (avg)
27.55 tok/s
Prefill (avg)
97.34 tok/s
TTFT (avg)
0.47 s
VRAM (avg)
67.3 GB
MTP acceptance rate
—
TTFB
0.22 GB
Power draw
377.3 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-27B
Quantization
Q8_0
Framework
vLLM
Version
v0.41
Context length
4096 tokens
Batch size
8
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 H20
Nominal VRAM
97 GB
Measured VRAM (avg)
67.3 GB
OS
Windows 11 24H2
Driver
560.94
CUDA
12.8
Power draw
377.3 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-19

    v0.41 · Windows 11 24H2 · CUDA 12.8 · 4096 ctx

    27.55 tok/s

    Decode

    97.34 tok/s

    Prefill

    0.47 s

    TTFT

    67.3 GB

    VRAM

    —

    MTP

    377.3 W

    Power draw

    L2 batch seed 2026-10-19: vLLM on hw_id=56 running model_id=50 at Q8_0