Arao AI
Back to search results

Qwen3-8B · FP16

Start a discussion
Apple M3 Max36GBvLLM v0.44L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-8B (FP16) on Apple M3 Max with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

57.8 tok/s

Decode

Decode speed

255.66 tok/s

Prefill

Prefill speed

0.91 s

TTFT

Time to first token

33.18 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. llama.cpp · FP16 Decode 58.36 · Prefill 290.84 ·
  2. llama.cpp · IQ2_M Decode 137.92 · Prefill 603.9 ·
  3. vLLM · FP16 (current)Decode 57.8 · Prefill 255.66 ·
  4. vLLM · IQ2_M Decode 145.31 · Prefill 613.3 ·
  5. Ollama · IQ2_M Decode 129.83 · Prefill 636.38 ·

Core figures

Decode (avg)
57.8 tok/s
Prefill (avg)
255.66 tok/s
TTFT (avg)
0.91 s
VRAM (avg)
33.18 GB
MTP acceptance rate
—
TTFB
0.46 GB
Power draw
42.4 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-8B
Quantization
FP16
Framework
vLLM
Version
v0.44
Context length
16384 tokens
Batch size
16
GPU layers
999
Flash Attention
Off

Hardware

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

GPU
Apple M3 Max
Nominal VRAM
36 GB
Measured VRAM (avg)
33.18 GB
OS
macOS 15.0
Driver
Apple M3 Max
CUDA
N/A
Power draw
42.4 W

Sources & evidence

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

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

    v0.44 · macOS 15.0 · CUDA N/A · 16384 ctx

    57.8 tok/s

    Decode

    255.66 tok/s

    Prefill

    0.91 s

    TTFT

    33.18 GB

    VRAM

    —

    MTP

    42.4 W

    Power draw

    L2 batch seed 2026-10-10: vLLM on hw_id=4 running model_id=53 at FP16