Arao AI
Back to search results

Qwen3.5-4B · IQ3_XS

Start a discussion
AMD Strix Halo (Radeon 8060S)Ollama v0.46L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-4B (IQ3_XS) on AMD Strix Halo (Radeon 8060S) with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

123.41 tok/s

Decode

Decode speed

403.1 tok/s

Prefill

Prefill speed

1.39 s

TTFT

Time to first token

7.31 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 · Q8_0 Decode 72.19 · Prefill 307.07 ·
  2. Ollama · IQ3_XS (current)Decode 123.41 · Prefill 403.1 ·

Core figures

Decode (avg)
123.41 tok/s
Prefill (avg)
403.1 tok/s
TTFT (avg)
1.39 s
VRAM (avg)
7.31 GB
MTP acceptance rate
—
TTFB
0.99 GB
Power draw
105.1 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-4B
Quantization
IQ3_XS
Framework
Ollama
Version
v0.46
Context length
4096 tokens
Batch size
256
GPU layers
33
Flash Attention
Off

Hardware

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

GPU
AMD Strix Halo (Radeon 8060S)
Measured VRAM (avg)
7.31 GB
OS
macOS 15.0
Driver
amdgpu 6.11
CUDA
N/A
Power draw
105.1 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 · 4096 ctx

    123.41 tok/s

    Decode

    403.1 tok/s

    Prefill

    1.39 s

    TTFT

    7.31 GB

    VRAM

    —

    MTP

    105.1 W

    Power draw

    L2 batch seed 2026-10-16: Ollama on hw_id=87 running model_id=48 at IQ3_XS