Arao AI
Back to search results

gemma-4-12B-it · Q4_K_S

Start a discussion
NVIDIA DGX Spark128GBStrata v0.43L2 Cross-framework verified

This page aggregates 1 real-world runs of gemma-4-12B-it (Q4_K_S) on NVIDIA DGX Spark with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

156.45 tok/s

Decode

Decode speed

581.38 tok/s

Prefill

Prefill speed

2.38 s

TTFT

Time to first token

21.76 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

  1. Ollama · Q2_K Decode 153.5 · Prefill 551.49 ·
  2. Strata · Q4_K_S (current)Decode 156.45 · Prefill 581.38 ·
  3. Strata · Q5_K_M Decode 139.24 · Prefill 652.49 ·

Core figures

Decode (avg)
156.45 tok/s
Prefill (avg)
581.38 tok/s
TTFT (avg)
2.38 s
VRAM (avg)
21.76 GB
MTP acceptance rate
0.63
TTFB
1.25 GB
Power draw
191.2 W

Configuration

Member-level fields are taken from the most recent run

Model
gemma-4-12B-it
Quantization
Q4_K_S
Framework
Strata
Version
v0.43
Context length
16384 tokens
Batch size
1024
GPU layers
12
Flash Attention
On

Hardware

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

GPU
NVIDIA DGX Spark
Nominal VRAM
128 GB
Measured VRAM (avg)
21.76 GB
OS
Ubuntu 24.04
Driver
560.94
CUDA
12.8
Power draw
191.2 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

    v0.43 · Ubuntu 24.04 · CUDA 12.8 · 16384 ctx

    156.45 tok/s

    Decode

    581.38 tok/s

    Prefill

    2.38 s

    TTFT

    21.76 GB

    VRAM

    0.63

    MTP

    191.2 W

    Power draw

    L2 batch seed 2026-10-13: Strata on hw_id=54 running model_id=65 at Q4_K_S