Arao AI
Back to search results

Qwen3.5-0.8B · FP16

Start a discussion
Apple M1 ProvLLM v0.46L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-0.8B (FP16) on Apple M1 Pro with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

88.25 tok/s

Decode

Decode speed

446.82 tok/s

Prefill

Prefill speed

0.63 s

TTFT

Time to first token

5.21 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

2 days ago

Last verified

Performance

  1. vLLM · FP16 (current)Decode 88.25 · Prefill 446.82 ·
  2. vLLM · IQ3_XS Decode 221.94 · Prefill 834.89 ·
  3. Strata · IQ4_XS Decode 263.91 · Prefill 1056.07 ·
  4. Strata · Q6_K Decode 165.63 · Prefill 531.23 ·

Core figures

Decode (avg)
88.25 tok/s
Prefill (avg)
446.82 tok/s
TTFT (avg)
0.63 s
VRAM (avg)
5.21 GB
MTP acceptance rate
—
TTFB
0.71 GB
Power draw
17.4 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-0.8B
Quantization
FP16
Framework
vLLM
Version
v0.46
Context length
2048 tokens
Batch size
16
GPU layers
999
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)
5.21 GB
OS
macOS 15.0
Driver
Apple M1 Pro
CUDA
N/A
Power draw
17.4 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-16

    v0.46 · macOS 15.0 · CUDA N/A · 2048 ctx

    88.25 tok/s

    Decode

    446.82 tok/s

    Prefill

    0.63 s

    TTFT

    5.21 GB

    VRAM

    —

    MTP

    17.4 W

    Power draw

    L2 batch seed 2026-10-16: vLLM on hw_id=100 running model_id=51 at FP16