Arao AI
Back to search results

gemma-4-31B-it · IQ4_XS

Start a discussion
Apple M5 Max128GBStrata v0.32L2 Cross-framework verified

This page aggregates 1 real-world runs of gemma-4-31B-it (IQ4_XS) on Apple M5 Max with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

138.4 tok/s

Decode

Decode speed

718.68 tok/s

Prefill

Prefill speed

1.19 s

TTFT

Time to first token

50.27 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

  1. Strata · IQ4_XS (current)Decode 138.4 · Prefill 718.68 ·
  2. Strata · Q3_K_M Decode 149.39 · Prefill 771.31 ·

Core figures

Decode (avg)
138.4 tok/s
Prefill (avg)
718.68 tok/s
TTFT (avg)
1.19 s
VRAM (avg)
50.27 GB
MTP acceptance rate
0.5
TTFB
1.11 GB
Power draw
36.9 W

Configuration

Member-level fields are taken from the most recent run

Model
gemma-4-31B-it
Quantization
IQ4_XS
Framework
Strata
Version
v0.32
Context length
8192 tokens
Batch size
256
GPU layers
5
Flash Attention
Off

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)
50.27 GB
OS
macOS 15.0
Driver
Apple M5 Max
CUDA
N/A
Power draw
36.9 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-13

    v0.32 · macOS 15.0 · CUDA N/A · 8192 ctx

    138.4 tok/s

    Decode

    718.68 tok/s

    Prefill

    1.19 s

    TTFT

    50.27 GB

    VRAM

    0.5

    MTP

    36.9 W

    Power draw

    L2 batch seed 2026-10-13: Strata on hw_id=28 running model_id=63 at IQ4_XS