Arao AI
Back to search results

Qwen3-4B · FP16

Start a discussion
AMD Strix Point (Radeon 890M)vLLM v0.33L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-4B (FP16) on AMD Strix Point (Radeon 890M) with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

36.08 tok/s

Decode

Decode speed

186.22 tok/s

Prefill

Prefill speed

1.09 s

TTFT

Time to first token

15.2 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 (current)Decode 36.08 · Prefill 186.22 ·
  2. vLLM · IQ2_XS 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)
36.08 tok/s
Prefill (avg)
186.22 tok/s
TTFT (avg)
1.09 s
VRAM (avg)
15.2 GB
MTP acceptance rate
—
TTFB
0.29 GB
Power draw
91.3 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-4B
Quantization
FP16
Framework
vLLM
Version
v0.33
Context length
32768 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)
15.2 GB
OS
Ubuntu 24.04
Driver
amdgpu 6.11
CUDA
N/A
Power draw
91.3 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

    v0.33 · Ubuntu 24.04 · CUDA N/A · 32768 ctx

    36.08 tok/s

    Decode

    186.22 tok/s

    Prefill

    1.09 s

    TTFT

    15.2 GB

    VRAM

    —

    MTP

    91.3 W

    Power draw

    L2 batch seed 2026-10-16: vLLM on hw_id=89 running model_id=54 at FP16