Arao AI
Back to search results

Qwen3-1.7B · Q6_K

Start a discussion
NVIDIA DGX Spark128GBllama.cpp v0.27L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-1.7B (Q6_K) on NVIDIA DGX Spark with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

210 tok/s

Decode

Decode speed

1,074.66 tok/s

Prefill

Prefill speed

0.23 s

TTFT

Time to first token

5.67 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

  1. vLLM · Q2_K Decode 374.76 · Prefill 1672.78 ·
  2. vLLM · Q5_K_M Decode 240.21 · Prefill 1012.49 ·
  3. Strata · IQ3_XS Decode 315.95 · Prefill 1231.57 ·
  4. Strata · Q4_K_S Decode 375.7 · Prefill 1600.4 ·
  5. llama.cpp · Q6_K (current)Decode 210 · Prefill 1074.66 ·
  6. vLLM · Q3_K_M Decode 331.74 · Prefill 1116.84 ·

Core figures

Decode (avg)
210 tok/s
Prefill (avg)
1,074.66 tok/s
TTFT (avg)
0.23 s
VRAM (avg)
5.67 GB
MTP acceptance rate
—
TTFB
1.68 GB
Power draw
268.3 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-1.7B
Quantization
Q6_K
Framework
llama.cpp
Version
v0.27
Context length
32768 tokens
Batch size
256
GPU layers
28
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)
5.67 GB
OS
Ubuntu 24.04
Driver
560.94
CUDA
12.8
Power draw
268.3 W

Sources & evidence

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

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

    v0.27 · Ubuntu 24.04 · CUDA 12.8 · 32768 ctx

    210 tok/s

    Decode

    1,074.66 tok/s

    Prefill

    0.23 s

    TTFT

    5.67 GB

    VRAM

    —

    MTP

    268.3 W

    Power draw

    L2 batch seed 2026-10-13: llama.cpp on hw_id=54 running model_id=55 at Q6_K