Arao AI
Back to search results

Qwen3-4B · IQ2_XS

Start a discussion
AMD Strix Point (Radeon 890M)vLLM v0.48L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-4B (IQ2_XS) on AMD Strix Point (Radeon 890M) with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

112.83 tok/s

Decode

Decode speed

404.8 tok/s

Prefill

Prefill speed

1.66 s

TTFT

Time to first token

6.13 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 · FP16 Decode 36.08 · Prefill 186.22 ·
  2. vLLM · IQ2_XS (current)Decode 112.83 · Prefill 404.8 ·
  3. Ollama · IQ4_XS Decode 68.56 · Prefill 288.08 ·
  4. Ollama · Q3_K_L Decode 75.1 · Prefill 314.07 ·

Core figures

Decode (avg)
112.83 tok/s
Prefill (avg)
404.8 tok/s
TTFT (avg)
1.66 s
VRAM (avg)
6.13 GB
MTP acceptance rate
—
TTFB
0.9 GB
Power draw
71.1 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-4B
Quantization
IQ2_XS
Framework
vLLM
Version
v0.48
Context length
8192 tokens
Batch size
16
GPU layers
999
Flash Attention
On

Hardware

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

GPU
AMD Strix Point (Radeon 890M)
Measured VRAM (avg)
6.13 GB
OS
macOS 15.0
Driver
amdgpu-pro 24.20
CUDA
N/A
Power draw
71.1 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

    v0.48 · macOS 15.0 · CUDA N/A · 8192 ctx

    112.83 tok/s

    Decode

    404.8 tok/s

    Prefill

    1.66 s

    TTFT

    6.13 GB

    VRAM

    —

    MTP

    71.1 W

    Power draw

    L2 batch seed 2026-10-16: vLLM on hw_id=89 running model_id=54 at IQ2_XS