Arao AI
Back to search results

gemma-4-E4B-it · Q5_K_S

Start a discussion
Apple M2 UltravLLM v0.34L2 Cross-framework verified

This page aggregates 1 real-world runs of gemma-4-E4B-it (Q5_K_S) on Apple M2 Ultra with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

123.25 tok/s

Decode

Decode speed

491.18 tok/s

Prefill

Prefill speed

1.83 s

TTFT

Time to first token

9.66 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

2 days ago

Last verified

Performance

  1. llama.cpp · Q3_K_S Decode 151.33 · Prefill 747.84 ·
  2. llama.cpp · Q5_K_S Decode 114.93 · Prefill 613.44 ·
  3. vLLM · Q5_K_S (current)Decode 123.25 · Prefill 491.18 ·

Core figures

Decode (avg)
123.25 tok/s
Prefill (avg)
491.18 tok/s
TTFT (avg)
1.83 s
VRAM (avg)
9.66 GB
MTP acceptance rate
—
TTFB
0.99 GB
Power draw
53.3 W

Configuration

Member-level fields are taken from the most recent run

Model
gemma-4-E4B-it
Quantization
Q5_K_S
Framework
vLLM
Version
v0.34
Context length
32768 tokens
Batch size
8
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 M2 Ultra
Measured VRAM (avg)
9.66 GB
OS
macOS 15.0
Driver
Apple M2 Ultra
CUDA
N/A
Power draw
53.3 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

    v0.34 · macOS 15.0 · CUDA N/A · 32768 ctx

    123.25 tok/s

    Decode

    491.18 tok/s

    Prefill

    1.83 s

    TTFT

    9.66 GB

    VRAM

    —

    MTP

    53.3 W

    Power draw

    L2 batch seed 2026-10-16: vLLM on hw_id=103 running model_id=67 at Q5_K_S