Arao AI
Back to search results

Qwen3.5-9B · Q5_K_M

Start a discussion
Apple M1 MaxStrata v0.28L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-9B (Q5_K_M) on Apple M1 Max with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

86.88 tok/s

Decode

Decode speed

306.63 tok/s

Prefill

Prefill speed

1.82 s

TTFT

Time to first token

19.68 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

2 days ago

Last verified

Performance

  1. llama.cpp · Q5_K_M Decode 65.43 · Prefill 334.37 ·
  2. Strata · FP16 Decode 50.07 · Prefill 196.15 ·
  3. Strata · Q5_K_M (current)Decode 86.88 · Prefill 306.63 ·

Core figures

Decode (avg)
86.88 tok/s
Prefill (avg)
306.63 tok/s
TTFT (avg)
1.82 s
VRAM (avg)
19.68 GB
MTP acceptance rate
0.72
TTFB
0.7 GB
Power draw
35.4 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-9B
Quantization
Q5_K_M
Framework
Strata
Version
v0.28
Context length
4096 tokens
Batch size
512
GPU layers
15
Flash Attention
On

Hardware

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

GPU
Apple M1 Max
Measured VRAM (avg)
19.68 GB
OS
macOS 15.0
Driver
Apple M1 Max
CUDA
N/A
Power draw
35.4 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-16

    v0.28 · macOS 15.0 · CUDA N/A · 4096 ctx

    86.88 tok/s

    Decode

    306.63 tok/s

    Prefill

    1.82 s

    TTFT

    19.68 GB

    VRAM

    0.72

    MTP

    35.4 W

    Power draw

    L2 batch seed 2026-10-16: Strata on hw_id=101 running model_id=49 at Q5_K_M