Arao AI
Back to search results

Qwen3-8B · IQ2_XS

Start a discussion
Intel Arc B58012GBvLLM v0.46L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-8B (IQ2_XS) on Intel Arc B580 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

78.59 tok/s

Decode

Decode speed

324.5 tok/s

Prefill

Prefill speed

1.04 s

TTFT

Time to first token

10.16 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. llama.cpp · Q3_K_M Decode 57.89 · Prefill 216.32 ·
  2. vLLM · IQ1_S Decode 82.24 · Prefill 450.81 ·
  3. vLLM · IQ2_XS (current)Decode 78.59 · Prefill 324.5 ·

Core figures

Decode (avg)
78.59 tok/s
Prefill (avg)
324.5 tok/s
TTFT (avg)
1.04 s
VRAM (avg)
10.16 GB
MTP acceptance rate
—
TTFB
0.63 GB
Power draw
258.8 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-8B
Quantization
IQ2_XS
Framework
vLLM
Version
v0.46
Context length
32768 tokens
Batch size
32
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
Intel Arc B580
Nominal VRAM
12 GB
Measured VRAM (avg)
10.16 GB
OS
Ubuntu 24.04
Driver
Intel Arc 31.0.101.5186
CUDA
12.6
Power draw
258.8 W

Sources & evidence

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

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

    v0.46 · Ubuntu 24.04 · CUDA 12.6 · 32768 ctx

    78.59 tok/s

    Decode

    324.5 tok/s

    Prefill

    1.04 s

    TTFT

    10.16 GB

    VRAM

    —

    MTP

    258.8 W

    Power draw

    L2 batch seed 2026-10-10: vLLM on hw_id=94 running model_id=53 at IQ2_XS