Arao AI
Back to search results

Qwen3.5-27B · IQ2_M

Start a discussion
Apple M1 ProOllama v0.31L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-27B (IQ2_M) on Apple M1 Pro with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

70.64 tok/s

Decode

Decode speed

337.05 tok/s

Prefill

Prefill speed

2 s

TTFT

Time to first token

30.4 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

2 days ago

Last verified

Performance

  1. Ollama · IQ2_M (current)Decode 70.64 · Prefill 337.05 ·

Core figures

Decode (avg)
70.64 tok/s
Prefill (avg)
337.05 tok/s
TTFT (avg)
2 s
VRAM (avg)
30.4 GB
MTP acceptance rate
—
TTFB
0.57 GB
Power draw
16.6 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-27B
Quantization
IQ2_M
Framework
Ollama
Version
v0.31
Context length
2048 tokens
Batch size
256
GPU layers
99
Flash Attention
On

Hardware

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

GPU
Apple M1 Pro
Measured VRAM (avg)
30.4 GB
OS
macOS 15.0
Driver
Apple M1 Pro
CUDA
N/A
Power draw
16.6 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

    v0.31 · macOS 15.0 · CUDA N/A · 2048 ctx

    70.64 tok/s

    Decode

    337.05 tok/s

    Prefill

    2 s

    TTFT

    30.4 GB

    VRAM

    —

    MTP

    16.6 W

    Power draw

    L2 batch seed 2026-10-16: Ollama on hw_id=100 running model_id=50 at IQ2_M