Arao AI
Back to search results

Qwen3.6-35B-A3B · IQ3_XS

Start a discussion
Apple M5 Max128GBllama.cpp v0.52L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.6-35B-A3B (IQ3_XS) on Apple M5 Max with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

159.32 tok/s

Decode

Decode speed

706.13 tok/s

Prefill

Prefill speed

0.4 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

17 days ago

Last verified

Performance

  1. llama.cpp · IQ3_XS (current)Decode 159.32 · Prefill 706.13 ·
  2. llama.cpp · Q6_K Decode 90.4 · Prefill 328.52 ·
  3. Ollama · Q3_K_L Decode 131.28 · Prefill 555.46 ·
  4. Ollama · Q5_K_S Decode 105.03 · Prefill 472.27 ·

Core figures

Decode (avg)
159.32 tok/s
Prefill (avg)
706.13 tok/s
TTFT (avg)
0.4 s
VRAM (avg)
47.67 GB
MTP acceptance rate
—
TTFB
1.27 GB
Power draw
57.1 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.6-35B-A3B
Quantization
IQ3_XS
Framework
llama.cpp
Version
v0.52
Context length
2048 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 M5 Max
Nominal VRAM
128 GB
Measured VRAM (avg)
47.67 GB
OS
macOS 26.1
Power draw
57.1 W

Sources & evidence

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

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

    v0.52 · macOS 26.1 · 2048 ctx

    159.32 tok/s

    Decode

    706.13 tok/s

    Prefill

    0.4 s

    TTFT

    47.67 GB

    VRAM

    —

    MTP

    57.1 W

    Power draw

    L2 batch seed 2026-10-19: llama.cpp on hw_id=28 running model_id=47 at IQ3_XS