Arao AI
Back to search results

Qwen3-Embedding-0.6B · IQ4_XS

Start a discussion
2x AMD Strix HaloOllama v0.46L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-Embedding-0.6B (IQ4_XS) on 2x AMD Strix Halo with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

379.81 tok/s

Decode

Decode speed

1,812.54 tok/s

Prefill

Prefill speed

1.53 s

TTFT

Time to first token

3.03 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

2 days ago

Last verified

Performance

  1. Ollama · FP16 Decode 181.91 · Prefill 736.26 ·
  2. Ollama · IQ4_XS (current)Decode 379.81 · Prefill 1812.54 ·
  3. Strata · IQ4_XS Decode 288.7 · Prefill 1574.67 ·

Core figures

Decode (avg)
379.81 tok/s
Prefill (avg)
1,812.54 tok/s
TTFT (avg)
1.53 s
VRAM (avg)
3.03 GB
MTP acceptance rate
—
TTFB
3.04 GB
Power draw
63.3 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-Embedding-0.6B
Quantization
IQ4_XS
Framework
Ollama
Version
v0.46
Context length
2048 tokens
Batch size
256
GPU layers
28
Flash Attention
Off

Hardware

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

GPU
2x AMD Strix Halo
Measured VRAM (avg)
3.03 GB
OS
macOS 15.0
Driver
570.86
CUDA
N/A
Power draw
63.3 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-16

    v0.46 · macOS 15.0 · CUDA N/A · 2048 ctx

    379.81 tok/s

    Decode

    1,812.54 tok/s

    Prefill

    1.53 s

    TTFT

    3.03 GB

    VRAM

    —

    MTP

    63.3 W

    Power draw

    L2 batch seed 2026-10-16: Ollama on hw_id=88 running model_id=62 at IQ4_XS