Arao AI
Back to search results

Qwen3-1.7B · Q3_K_M

Start a discussion
Intel Arc B58012GBvLLM v0.31L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-1.7B (Q3_K_M) on Intel Arc B580 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

109.13 tok/s

Decode

Decode speed

478.21 tok/s

Prefill

Prefill speed

0.47 s

TTFT

Time to first token

4.53 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. llama.cpp · IQ2_XS Decode 140.48 · Prefill 573.63 ·
  2. llama.cpp · Q6_K Decode 72.12 · Prefill 270.78 ·
  3. vLLM · IQ3_S Decode 115.58 · Prefill 443.88 ·
  4. vLLM · Q3_K_M (current)Decode 109.13 · Prefill 478.21 ·
  5. Strata · IQ3_S Decode 143.13 · Prefill 545.08 ·

Core figures

Decode (avg)
109.13 tok/s
Prefill (avg)
478.21 tok/s
TTFT (avg)
0.47 s
VRAM (avg)
4.53 GB
MTP acceptance rate
—
TTFB
0.87 GB
Power draw
199.1 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-1.7B
Quantization
Q3_K_M
Framework
vLLM
Version
v0.31
Context length
32768 tokens
Batch size
8
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
Intel Arc B580
Nominal VRAM
12 GB
Measured VRAM (avg)
4.53 GB
OS
Ubuntu 24.04
Driver
Intel Arc 31.0.101.5186
CUDA
12.6
Power draw
199.1 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

    v0.31 · Ubuntu 24.04 · CUDA 12.6 · 32768 ctx

    109.13 tok/s

    Decode

    478.21 tok/s

    Prefill

    0.47 s

    TTFT

    4.53 GB

    VRAM

    —

    MTP

    199.1 W

    Power draw

    L2 batch seed 2026-10-10: vLLM on hw_id=94 running model_id=55 at Q3_K_M