Arao AI
Back to search results

Qwen3.6-27B · Q3_K_L

Start a discussion
Apple M2 UltraOllama v0.40L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.6-27B (Q3_K_L) on Apple M2 Ultra with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

80.85 tok/s

Decode

Decode speed

325.8 tok/s

Prefill

Prefill speed

1.44 s

TTFT

Time to first token

42.92 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 · Q4_K_M Decode 67.38 · Prefill 344.18 ·
  2. Ollama · Q3_K_L (current)Decode 80.85 · Prefill 325.8 ·

Core figures

Decode (avg)
80.85 tok/s
Prefill (avg)
325.8 tok/s
TTFT (avg)
1.44 s
VRAM (avg)
42.92 GB
MTP acceptance rate
—
TTFB
0.65 GB
Power draw
109.7 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.6-27B
Quantization
Q3_K_L
Framework
Ollama
Version
v0.40
Context length
8192 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 M2 Ultra
Measured VRAM (avg)
42.92 GB
OS
macOS 26.1
Power draw
109.7 W

Get started

Original model resources

Base model resources may not match this quantization. Use the verified deployment weights above when available.

Sources & evidence

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

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

    v0.40 · macOS 26.1 · 8192 ctx

    80.85 tok/s

    Decode

    325.8 tok/s

    Prefill

    1.44 s

    TTFT

    42.92 GB

    VRAM

    —

    MTP

    109.7 W

    Power draw

    L2 batch seed 2026-10-19: Ollama on hw_id=103 running model_id=6 at Q3_K_L