Arao AI
Back to search results

gemma-4-31B-it · Q6_K

Start a discussion
NVIDIA DGX Spark128GBOllama v0.37L2 Cross-framework verified

This page aggregates 1 real-world runs of gemma-4-31B-it (Q6_K) on NVIDIA DGX Spark with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

50.14 tok/s

Decode

Decode speed

188.84 tok/s

Prefill

Prefill speed

0.23 s

TTFT

Time to first token

67.2 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

  1. llama.cpp · IQ3_XS Decode 91.61 · Prefill 449.28 ·
  2. Ollama · Q6_K (current)Decode 50.14 · Prefill 188.84 ·

Core figures

Decode (avg)
50.14 tok/s
Prefill (avg)
188.84 tok/s
TTFT (avg)
0.23 s
VRAM (avg)
67.2 GB
MTP acceptance rate
—
TTFB
0.4 GB
Power draw
196.9 W

Configuration

Member-level fields are taken from the most recent run

Model
gemma-4-31B-it
Quantization
Q6_K
Framework
Ollama
Version
v0.37
Context length
2048 tokens
Batch size
256
GPU layers
20
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)
67.2 GB
OS
Ubuntu 24.04
Driver
560.94
CUDA
12.8
Power draw
196.9 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

    v0.37 · Ubuntu 24.04 · CUDA 12.8 · 2048 ctx

    50.14 tok/s

    Decode

    188.84 tok/s

    Prefill

    0.23 s

    TTFT

    67.2 GB

    VRAM

    —

    MTP

    196.9 W

    Power draw

    L2 batch seed 2026-10-13: Ollama on hw_id=54 running model_id=63 at Q6_K