Arao AI
Back to search results

Qwen3.6-35B-A3B · IQ3_XS

Start a discussion
Apple M4 Max128GBllama.cpp v0.23L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.6-35B-A3B (IQ3_XS) on Apple M4 Max with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

121.63 tok/s

Decode

Decode speed

540.05 tok/s

Prefill

Prefill speed

2.37 s

TTFT

Time to first token

47.67 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. llama.cpp · IQ3_XS (current)Decode 121.63 · Prefill 540.05 ·
  2. llama.cpp · Q4_K_S Decode 111.01 · Prefill 387.61 ·
  3. Ollama · Q3_K_L Decode 102.3 · Prefill 395.71 ·

Core figures

Decode (avg)
121.63 tok/s
Prefill (avg)
540.05 tok/s
TTFT (avg)
2.37 s
VRAM (avg)
47.67 GB
MTP acceptance rate
—
TTFB
0.97 GB
Power draw
63 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.6-35B-A3B
Quantization
IQ3_XS
Framework
llama.cpp
Version
v0.23
Context length
8192 tokens
Batch size
512
GPU layers
99
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)
47.67 GB
OS
macOS 15.0
Driver
Apple M4 Max
CUDA
N/A
Power draw
63 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

    v0.23 · macOS 15.0 · CUDA N/A · 8192 ctx

    121.63 tok/s

    Decode

    540.05 tok/s

    Prefill

    2.37 s

    TTFT

    47.67 GB

    VRAM

    —

    MTP

    63 W

    Power draw

    L2 batch seed 2026-10-10: llama.cpp on hw_id=26 running model_id=47 at IQ3_XS