Arao AI
Back to search results

Qwen3.8-9B · IQ3_S

Start a discussion
Apple M4 Max128GBStrata v0.51L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.8-9B (IQ3_S) on Apple M4 Max with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

255.89 tok/s

Decode

Decode speed

1,167.45 tok/s

Prefill

Prefill speed

0.47 s

TTFT

Time to first token

14.57 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. vLLM · IQ2_XS Decode 251.74 · Prefill 1248.2 ·
  2. Ollama · Q8_0 Decode 103.18 · Prefill 552.91 ·
  3. Strata · IQ2_XS Decode 353.9 · Prefill 1746.97 ·
  4. Strata · IQ3_S (current)Decode 255.89 · Prefill 1167.45 ·

Core figures

Decode (avg)
255.89 tok/s
Prefill (avg)
1,167.45 tok/s
TTFT (avg)
0.47 s
VRAM (avg)
14.57 GB
MTP acceptance rate
0.57
TTFB
2.05 GB
Power draw
52 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.8-9B
Quantization
IQ3_S
Framework
Strata
Version
v0.51
Context length
2048 tokens
Batch size
512
GPU layers
5
Flash Attention
Off

Hardware

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

GPU
Apple M4 Max
Nominal VRAM
128 GB
Measured VRAM (avg)
14.57 GB
OS
macOS 15.0
Driver
Apple M4 Max
CUDA
N/A
Power draw
52 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

    v0.51 · macOS 15.0 · CUDA N/A · 2048 ctx

    255.89 tok/s

    Decode

    1,167.45 tok/s

    Prefill

    0.47 s

    TTFT

    14.57 GB

    VRAM

    0.57

    MTP

    52 W

    Power draw

    L2 batch seed 2026-10-10: Strata on hw_id=26 running model_id=52 at IQ3_S