Arao AI
Back to search results

Llama-3-8B-Instruct · IQ3_XS

Start a discussion
Apple M5 Ultra512GBOllama v0.42L2 Cross-framework verified

This page aggregates 1 real-world runs of Llama-3-8B-Instruct (IQ3_XS) on Apple M5 Ultra with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

140.68 tok/s

Decode

Decode speed

755.32 tok/s

Prefill

Prefill speed

0.33 s

TTFT

Time to first token

12.52 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

16 days ago

Last verified

Performance

  1. llama.cpp · IQ2_M Decode 145.48 · Prefill 672.26 ·
  2. llama.cpp · Q4_K_S Decode 115.5 · Prefill 509.11 ·
  3. Ollama · IQ3_XS (current)Decode 140.68 · Prefill 755.32 ·

Core figures

Decode (avg)
140.68 tok/s
Prefill (avg)
755.32 tok/s
TTFT (avg)
0.33 s
VRAM (avg)
12.52 GB
MTP acceptance rate
—
TTFB
1.13 GB
Power draw
90.8 W

Configuration

Member-level fields are taken from the most recent run

Model
Llama-3-8B-Instruct
Quantization
IQ3_XS
Framework
Ollama
Version
v0.42
Context length
2048 tokens
Batch size
512
GPU layers
33
Flash Attention
On

Hardware

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

GPU
Apple M5 Ultra
Nominal VRAM
512 GB
Measured VRAM (avg)
12.52 GB
OS
macOS 26.1
Power draw
90.8 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-19

    v0.42 · macOS 26.1 · 2048 ctx

    140.68 tok/s

    Decode

    755.32 tok/s

    Prefill

    0.33 s

    TTFT

    12.52 GB

    VRAM

    —

    MTP

    90.8 W

    Power draw

    L2 batch seed 2026-10-19: Ollama on hw_id=29 running model_id=69 at IQ3_XS