Arao AI
Back to search results

Qwen3.8-9B · IQ3_XS

Start a discussion
Apple M2 UltraOllama v0.37L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.8-9B (IQ3_XS) on Apple M2 Ultra with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

108.03 tok/s

Decode

Decode speed

420.83 tok/s

Prefill

Prefill speed

0.13 s

TTFT

Time to first token

13.82 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

17 days ago

Last verified

Performance

  1. llama.cpp · Q3_K_L Decode 110.22 · Prefill 411.86 ·
  2. Ollama · IQ3_XS (current)Decode 108.03 · Prefill 420.83 ·
  3. vLLM · IQ1_S Decode 152.06 · Prefill 546.19 ·
  4. Strata · Q4_K_M Decode 129.12 · Prefill 454.27 ·
  5. Strata · Q8_0 Decode 89.33 · Prefill 303.19 ·

Core figures

Decode (avg)
108.03 tok/s
Prefill (avg)
420.83 tok/s
TTFT (avg)
0.13 s
VRAM (avg)
13.82 GB
MTP acceptance rate
—
TTFB
0.86 GB
Power draw
141.3 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.8-9B
Quantization
IQ3_XS
Framework
Ollama
Version
v0.37
Context length
8192 tokens
Batch size
512
GPU layers
33
Flash Attention
Off

Hardware

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

GPU
Apple M2 Ultra
Measured VRAM (avg)
13.82 GB
OS
macOS 26.1
Power draw
141.3 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

    v0.37 · macOS 26.1 · 8192 ctx

    108.03 tok/s

    Decode

    420.83 tok/s

    Prefill

    0.13 s

    TTFT

    13.82 GB

    VRAM

    —

    MTP

    141.3 W

    Power draw

    L2 batch seed 2026-10-19: Ollama on hw_id=103 running model_id=52 at IQ3_XS