Arao AI
Back to search results

Qwen3-1.7B · IQ3_XS

Start a discussion
NVIDIA DGX Spark128GBStrata v0.39L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-1.7B (IQ3_XS) on NVIDIA DGX Spark with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

315.95 tok/s

Decode

Decode speed

1,231.57 tok/s

Prefill

Prefill speed

2.12 s

TTFT

Time to first token

4.32 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

2 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 (current)Decode 315.95 · Prefill 1231.57 ·
  4. Strata · Q4_K_S Decode 375.7 · Prefill 1600.4 ·
  5. llama.cpp · Q6_K Decode 210 · Prefill 1074.66 ·
  6. vLLM · Q3_K_M Decode 331.74 · Prefill 1116.84 ·

Core figures

Decode (avg)
315.95 tok/s
Prefill (avg)
1,231.57 tok/s
TTFT (avg)
2.12 s
VRAM (avg)
4.32 GB
MTP acceptance rate
0.67
TTFB
2.53 GB
Power draw
256.6 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-1.7B
Quantization
IQ3_XS
Framework
Strata
Version
v0.39
Context length
8192 tokens
Batch size
256
GPU layers
8
Flash Attention
Off

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)
4.32 GB
OS
Windows 11 23H2
Driver
555.85
CUDA
12.6
Power draw
256.6 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

    v0.39 · Windows 11 23H2 · CUDA 12.6 · 8192 ctx

    315.95 tok/s

    Decode

    1,231.57 tok/s

    Prefill

    2.12 s

    TTFT

    4.32 GB

    VRAM

    0.67

    MTP

    256.6 W

    Power draw

    L2 batch seed 2026-10-16: Strata on hw_id=54 running model_id=55 at IQ3_XS