Arao AI
Back to search results

Qwen3.8-9B · IQ4_XS

Start a discussion
NVIDIA RTX 409024GBvLLM v0.36L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.8-9B (IQ4_XS) on NVIDIA RTX 4090 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

131.27 tok/s

Decode

Decode speed

643.79 tok/s

Prefill

Prefill speed

0.96 s

TTFT

Time to first token

16.09 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. vLLM · IQ4_XS (current)Decode 131.27 · Prefill 643.79 ·
  2. vLLM · Q2_K Decode 152.53 · Prefill 504.65 ·
  3. Ollama · IQ3_XS Decode 121.23 · Prefill 396.17 ·

Core figures

Decode (avg)
131.27 tok/s
Prefill (avg)
643.79 tok/s
TTFT (avg)
0.96 s
VRAM (avg)
16.09 GB
MTP acceptance rate
—
TTFB
1.05 GB
Power draw
322.9 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.8-9B
Quantization
IQ4_XS
Framework
vLLM
Version
v0.36
Context length
32768 tokens
Batch size
16
GPU layers
999
Flash Attention
Off

Hardware

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

GPU
NVIDIA RTX 4090
Nominal VRAM
24 GB
Measured VRAM (avg)
16.09 GB
OS
Ubuntu 24.04
Driver
572.83
CUDA
12.8
Power draw
322.9 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-18

    v0.36 · Ubuntu 24.04 · CUDA 12.8 · 32768 ctx

    131.27 tok/s

    Decode

    643.79 tok/s

    Prefill

    0.96 s

    TTFT

    16.09 GB

    VRAM

    —

    MTP

    322.9 W

    Power draw

    L2 batch seed 2026-10-18: vLLM on hw_id=1 running model_id=52 at IQ4_XS