Arao AI
Back to search results

Llama-3.2-3B-Instruct · FP16

Start a discussion
Intel Arc B58012GBStrata v0.38L2 Cross-framework verified

This page aggregates 1 real-world runs of Llama-3.2-3B-Instruct (FP16) on Intel Arc B580 with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

92 tok/s

Decode

Decode speed

368.95 tok/s

Prefill

Prefill speed

0.5 s

TTFT

Time to first token

11.4 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. vLLM · IQ3_XS Decode 157.7 · Prefill 718.89 ·
  2. vLLM · Q4_K_M Decode 135.38 · Prefill 477.15 ·
  3. Strata · FP16 (current)Decode 92 · Prefill 368.95 ·
  4. Strata · Q4_K_M Decode 163.66 · Prefill 710.01 ·

Core figures

Decode (avg)
92 tok/s
Prefill (avg)
368.95 tok/s
TTFT (avg)
0.5 s
VRAM (avg)
11.4 GB
MTP acceptance rate
0.71
TTFB
0.74 GB
Power draw
162.8 W

Configuration

Member-level fields are taken from the most recent run

Model
Llama-3.2-3B-Instruct
Quantization
FP16
Framework
Strata
Version
v0.38
Context length
2048 tokens
Batch size
512
GPU layers
15
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)
11.4 GB
OS
Windows 11 24H2
Driver
Intel Arc 32.0.101.6552
CUDA
12.6
Power draw
162.8 W

Sources & evidence

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

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

    v0.38 · Windows 11 24H2 · CUDA 12.6 · 2048 ctx

    92 tok/s

    Decode

    368.95 tok/s

    Prefill

    0.5 s

    TTFT

    11.4 GB

    VRAM

    0.71

    MTP

    162.8 W

    Power draw

    L2 batch seed 2026-10-18: Strata on hw_id=94 running model_id=70 at FP16