Arao AI
Back to search results

Llama-3-8B-Instruct · IQ1_S

Start a discussion
Intel Arc B58012GBvLLM v0.48L2 Cross-framework verified

This page aggregates 1 real-world runs of Llama-3-8B-Instruct (IQ1_S) on Intel Arc B580 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

85.09 tok/s

Decode

Decode speed

376.72 tok/s

Prefill

Prefill speed

1.85 s

TTFT

Time to first token

9.16 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. llama.cpp · IQ3_S Decode 60.72 · Prefill 260.85 ·
  2. vLLM · IQ1_S (current)Decode 85.09 · Prefill 376.72 ·
  3. vLLM · Q2_K Decode 76.78 · Prefill 352.63 ·

Core figures

Decode (avg)
85.09 tok/s
Prefill (avg)
376.72 tok/s
TTFT (avg)
1.85 s
VRAM (avg)
9.16 GB
MTP acceptance rate
—
TTFB
0.68 GB
Power draw
153.5 W

Configuration

Member-level fields are taken from the most recent run

Model
Llama-3-8B-Instruct
Quantization
IQ1_S
Framework
vLLM
Version
v0.48
Context length
4096 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)
9.16 GB
OS
Ubuntu 24.04
Driver
Intel Arc 32.0.101.6552
CUDA
12.6
Power draw
153.5 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

    v0.48 · Ubuntu 24.04 · CUDA 12.6 · 4096 ctx

    85.09 tok/s

    Decode

    376.72 tok/s

    Prefill

    1.85 s

    TTFT

    9.16 GB

    VRAM

    —

    MTP

    153.5 W

    Power draw

    L2 batch seed 2026-10-18: vLLM on hw_id=94 running model_id=69 at IQ1_S